Pinned Tweet
EKOS Technical Progress & Tokenomics 2.0
One engineer. 100 days. 508 commits.
And we’re just getting started. 🚀
Could your software engineers work this effectively?
I ran ZAUTH on my open-source project, EKOS, for the second time.
📊 Aug 23 → Oct 19
• 126 → 508 commits
• ZAro external code matches reported
• No high-risk security patterns flagged
4x the commits. Real engineering progress.
Here’s what sits behind those numbers:
🧠 Original Engineering
• fact_ledger.rs: A custom log-structured storage engine built for time-travel queries.
• semantics.rs: Business semantics synthesis with Git-blame-based hypothesis tracking.
• cross_system.rs: Cross-database identity resolution using Jaro-Winkler and Jaccard scoring.
🛡 Built with Safeguards
• Consistent Result<T> error handling
• Token validation on the setup server
• Defensive coding practices designed to reduce common attack vectors
✅ Tested Where It Matters
• Unit and integration tests across major modules
• Time-travel edge cases and cross-database identity collisions
• AI agents using MCP cannot resolve conflicts autonomously. That decision stays with humans.
📐 Architecture Before Implementation
• RFC-driven development
• A modular Rust workspace spanning ledger, identity, CLI, EKL, and distributed components
This isn’t about chasing commit counts. It’s about building infrastructure engineers can actually use to bridge legacy systems and AI.
And this is where solana:CwubepDFJndzSKFmAMAm9u8Xx3PrizAwSq8hcGimpump becomes part of the bigger picture.
We’re building the technology first. At the same time, we’re exploring how the ecosystem can better recognize and reward meaningful contributions—from development and testing to community participation.
That’s the thinking behind the ongoing Tokenomics 2.0 discussions: aligning the long-term interests of builders, contributors, and the wider EKOS community.
The goal isn’t just to launch a token around an idea.
It’s to build something real—and develop an ecosystem worthy of growing around it.
More engineering. More transparency. More opportunities for the community to contribute.
EKOS is open source. We’re building in public. And we’re far from done. 🔥
🔗 ZAUTH: zauth.inc/reposcan/scan/…
💻 GitHub: github.com/alexeyban/EKOS/
solana:CwubepDFJndzSKFmAMAm9u8Xx3PrizAwSq8hcGimpump #Rust #OpenSource #AI #DataEngineering
@IBM @Dell @intel @AnthropicAI @UberEng @micro1_ai @pentagoniac @databricks
One engineer. 100 days. 508 commits.
And we’re just getting started. 🚀
Could your software engineers work this effectively?
I ran ZAUTH on my open-source project, EKOS, for the second time.
📊 Aug 23 → Oct 19
• 126 → 508 commits
• ZAro external code matches reported
• No high-risk security patterns flagged
4x the commits. Real engineering progress.
Here’s what sits behind those numbers:
🧠 Original Engineering
• fact_ledger.rs: A custom log-structured storage engine built for time-travel queries.
• semantics.rs: Business semantics synthesis with Git-blame-based hypothesis tracking.
• cross_system.rs: Cross-database identity resolution using Jaro-Winkler and Jaccard scoring.
🛡 Built with Safeguards
• Consistent Result<T> error handling
• Token validation on the setup server
• Defensive coding practices designed to reduce common attack vectors
✅ Tested Where It Matters
• Unit and integration tests across major modules
• Time-travel edge cases and cross-database identity collisions
• AI agents using MCP cannot resolve conflicts autonomously. That decision stays with humans.
📐 Architecture Before Implementation
• RFC-driven development
• A modular Rust workspace spanning ledger, identity, CLI, EKL, and distributed components
This isn’t about chasing commit counts. It’s about building infrastructure engineers can actually use to bridge legacy systems and AI.
And this is where solana:CwubepDFJndzSKFmAMAm9u8Xx3PrizAwSq8hcGimpump becomes part of the bigger picture.
We’re building the technology first. At the same time, we’re exploring how the ecosystem can better recognize and reward meaningful contributions—from development and testing to community participation.
That’s the thinking behind the ongoing Tokenomics 2.0 discussions: aligning the long-term interests of builders, contributors, and the wider EKOS community.
The goal isn’t just to launch a token around an idea.
It’s to build something real—and develop an ecosystem worthy of growing around it.
More engineering. More transparency. More opportunities for the community to contribute.
EKOS is open source. We’re building in public. And we’re far from done. 🔥
🔗 ZAUTH: zauth.inc/reposcan/scan/…
💻 GitHub: github.com/alexeyban/EKOS/
solana:CwubepDFJndzSKFmAMAm9u8Xx3PrizAwSq8hcGimpump #Rust #OpenSource #AI #DataEngineering
@IBM @Dell @intel @AnthropicAI @UberEng @micro1_ai @pentagoniac @databricks
21
14
58
7,656
🚀 WE’RE BACK ON X!
Thanks to our team’s hard work, we’re back and ready to move forward!
First, a huge thank-you to everyone who supported us, stayed patient, and kept believing in what we’re building. Your support means a lot to us. ❤️
Going forward, we’ll make our posts gic while ensuring that important product updates and development progress continue to reach you without interruption.
🔥 And this is just the beginning…
We have some major news to share today. A lot has been happening behind the scenes, and we can’t wait to show you what’s next.
EKOS keeps building. EKOS keeps moving forward.
Stay tuned. 👀
Thanks to our team’s hard work, we’re back and ready to move forward!
First, a huge thank-you to everyone who supported us, stayed patient, and kept believing in what we’re building. Your support means a lot to us. ❤️
Going forward, we’ll make our posts gic while ensuring that important product updates and development progress continue to reach you without interruption.
🔥 And this is just the beginning…
We have some major news to share today. A lot has been happening behind the scenes, and we can’t wait to show you what’s next.
EKOS keeps building. EKOS keeps moving forward.
Stay tuned. 👀
5
4
20
118
Ekos Project retweeted
Beyond the demo: Discover how EKOS tackles real-world complexity, enterprise-scale data, and practical engineering challenges. x.com/i/article/2114…
2
2
10
254
🔥 Something Real Is Coming.
Not another AI promise. Not just a pipeline diagram.
A real Oracle-to-Lakehouse workflow.
Real data. Real validation. Real issues found and fixed.
33 tables. 895 checks executed. Three Gold marts reconciled against Silver.
And we caught two real bugs while recording the demo — then fixed them.
This is how we move toward zero errors: test, verify, fix, repeat.
Full end-to-end demo coming soon.
#EKOS #DataEngineering
github.com/alexeyban/EKOS…
Not another AI promise. Not just a pipeline diagram.
A real Oracle-to-Lakehouse workflow.
Real data. Real validation. Real issues found and fixed.
33 tables. 895 checks executed. Three Gold marts reconciled against Silver.
And we caught two real bugs while recording the demo — then fixed them.
This is how we move toward zero errors: test, verify, fix, repeat.
Full end-to-end demo coming soon.
#EKOS #DataEngineering
github.com/alexeyban/EKOS…
1
4
21
263
Ekos Project retweeted
Latest News - solana:CwubepDFJndzSKFmAMAm9u8Xx3PrizAwSq8hcGimpump
@ekosproject has completed a major test of compiling legacy ERP knowledge into an evidence-backed lakehouse.
On iDempiere, an open-source Java ERP running on Oracle, it analyzed 4,542 Java files, generated 1,060 pages of source-system documentation, tested 165 tables end-to-end with every Bronze table reconciled against Oracle, and saw 3,664 of 3,667 Silver tests pass while uncovering 6 real bugs, all with zero LLM calls from analysis through Silver. EKOS compiled source code, database migrations, Git history, documentation, and the live Oracle catalog into a knowledge ledger that informed a Trino-to-Iceberg Bronze, dbt-trino Silver, and Dagster-orchestrated pipeline, with incremental loading and delete detection drawn from source-code behavior.
EKOS is an open-source Knowledge Compiler for AI agents, built in Rust with MCP. The full 849-table run remains memory-constrained and 195 conflicts still await human review, while Gold Mart drafting continues. The work moves legacy modernization toward traceable, testable pipelines grounded in how the enterprise system actually behaves.
For all the latest Solana tech news, check out techterminal.info/news. We track and cover updates from over 200+ Solana tech projects in one place.
@ekosproject has completed a major test of compiling legacy ERP knowledge into an evidence-backed lakehouse.
On iDempiere, an open-source Java ERP running on Oracle, it analyzed 4,542 Java files, generated 1,060 pages of source-system documentation, tested 165 tables end-to-end with every Bronze table reconciled against Oracle, and saw 3,664 of 3,667 Silver tests pass while uncovering 6 real bugs, all with zero LLM calls from analysis through Silver. EKOS compiled source code, database migrations, Git history, documentation, and the live Oracle catalog into a knowledge ledger that informed a Trino-to-Iceberg Bronze, dbt-trino Silver, and Dagster-orchestrated pipeline, with incremental loading and delete detection drawn from source-code behavior.
EKOS is an open-source Knowledge Compiler for AI agents, built in Rust with MCP. The full 849-table run remains memory-constrained and 195 conflicts still await human review, while Gold Mart drafting continues. The work moves legacy modernization toward traceable, testable pipelines grounded in how the enterprise system actually behaves.
For all the latest Solana tech news, check out techterminal.info/news. We track and cover updates from over 200+ Solana tech projects in one place.
🚀 A MAJOR MILESTONE FOR EKOS 🔥
From legacy ERP source code to a reconciled Lakehouse.
What happens when enterprise data pipelines are built on evidence from the source system—not assumptions?
We put EKOS to the test on iDempiere, an open-source Java ERP running on Oracle.
🔥 THE RESULTS:
🔍 4,542 Java files analyzed
📚 1,060 pages of source-system documentation generated
🧪 165 tables tested end-to-end, with every Bronze table reconciled against Oracle
📊 3,664 of 3,667 Silver tests passed
🛠️ 6 real bugs uncovered during testing
⚙️ Zero LLM calls from analysis through Silver
But the real milestone goes beyond the numbers.
EKOS compiled source code, database migrations, Git history, documentation, and the live Oracle catalog into an evidence-backed knowledge ledger.
That evidence informed the pipeline:
→ Trino → Iceberg Bronze
→ dbt-trino Silver
→ Dagster orchestration
Incremental loading is selected based on source-code behavior. Deletes are detected through bucketed key comparisons, with safeguards that hold mass deletions for human approval. Silver tests derive from database constraints and the Java model.
Every pipeline decision carries its evidence.
The full 849-table run remains constrained by available memory, and 195 conflicts still await human review. Gold Mart drafting works; the remaining work is ongoing.
This is a major step toward making legacy modernization more traceable, testable, and grounded in how enterprise systems actually behave.
EKOS: Enterprise knowledge, compiled into engineering evidence.
#EKOS #DataEngineering #Lakehouse #OpenSource4o
@Oracle @OracleCloud @googlecloud @thatjeffsmith @SQLServer @Dell @IBM
2
4
19
394
Ekos Project retweeted
@OMID_0909 @suraj_sharma14 The structure you describe falls out naturally if memory is an event log: context, retrieval and observability are all projections of the same history instead of three separate systems. That is how AllSource Prime is built. Happy to compare notes.
1
3
5
110
Ekos Project retweeted
EKOS is getting noticed, and the response has been amazing! 🔥
We’re getting ready to test it with bigger datasets from major companies and share the results soon.
The product is getting the attention it deserves. We’re just getting started! 🚀
Soon, you’ll see some big names @ekosproject …….
We’re getting ready to test it with bigger datasets from major companies and share the results soon.
The product is getting the attention it deserves. We’re just getting started! 🚀
Soon, you’ll see some big names @ekosproject …….
10
8
32
840
Ekos Project retweeted
Holy f*ck, Google just released a free 1-hour course on agentic knowledge graphs
"Your agents don't need a bigger model. They need a graph that remembers everything"
Prompts → Agents → Knowledge Graphs → Self-Improving Agents
04:01 - Build your first GraphRAG agent from search: agents that find the answer on their own
35:48 - Graph engineering in practice, live
47:06 - Self-improving agents that get smarter with every run
While you scroll, someone else is learning Google's $750,000 skill
This free course is better than any $1,500 graph engineering bootcamp
Bookmark and watch it today
Then read the full harness engineering guide below ↓
"Your agents don't need a bigger model. They need a graph that remembers everything"
Prompts → Agents → Knowledge Graphs → Self-Improving Agents
04:01 - Build your first GraphRAG agent from search: agents that find the answer on their own
35:48 - Graph engineering in practice, live
47:06 - Self-improving agents that get smarter with every run
While you scroll, someone else is learning Google's $750,000 skill
This free course is better than any $1,500 graph engineering bootcamp
Bookmark and watch it today
Then read the full harness engineering guide below ↓
4
54
379
17,975
EKOS UPDATE | Stronger Validation. Greater Reliability. 🔥
Moving enterprise data is not enough. Every transformation, every change, and every critical operation must be validated.
⚙️🛠️👷♂️
That’s what we’re building into EKOS:
🔹 Traceability: Turning real execution steps into structured reports and evidence-backed documentation.
🔹 Performance & security: Strengthening software dependencies while improving search performance, with a reported 41% speedup in the find_objects benchmark.
🔹 Validation & control: Expanding release-readiness checks, identifying data-level risks, and tracking unresolved issues instead of letting them go unnoticed.
Not just moving data. Making every step verifiable.
EKOS — Driving Errors Toward Zero. 🔥
#EKOS #EnterpriseAI #DataEngineering
@googlecloud @IBM @Dell @intel @Microsoft @awscloud
@stripe @UberEng @AnthropicAI
@OpenAI @elonmusk
🔗 Explore the latest validation and release-readiness work:
github.com/alexeyban/EKOS…
Moving enterprise data is not enough. Every transformation, every change, and every critical operation must be validated.
⚙️🛠️👷♂️
That’s what we’re building into EKOS:
🔹 Traceability: Turning real execution steps into structured reports and evidence-backed documentation.
🔹 Performance & security: Strengthening software dependencies while improving search performance, with a reported 41% speedup in the find_objects benchmark.
🔹 Validation & control: Expanding release-readiness checks, identifying data-level risks, and tracking unresolved issues instead of letting them go unnoticed.
Not just moving data. Making every step verifiable.
EKOS — Driving Errors Toward Zero. 🔥
#EKOS #EnterpriseAI #DataEngineering
@googlecloud @IBM @Dell @intel @Microsoft @awscloud
@stripe @UberEng @AnthropicAI
@OpenAI @elonmusk
🔗 Explore the latest validation and release-readiness work:
github.com/alexeyban/EKOS…
8
9
33
479
⚙️ We’re working relentlessly to deliver a product with errors as close to zero as possible.
This is more than progress. This is the future. 🚀🔥
@ekosproject is future ⚙️
This is more than progress. This is the future. 🚀🔥
@ekosproject is future ⚙️
2
0
7
138
Ekos Project retweeted
🚀 The numbers tell the story. Now, it's time to SHOW the world.
EKOS has delivered measurable results in its legacy-to-Lakehouse test. But we're not stopping at numbers on a screen.
🎬 We're bringing the entire process to life.
From analyzing thousands of Java files to ainst Oracle, we're preparing a full demo that puts the results front and center.
Code tells the truth. Seeing it happen makes that truth impossible to ignore.
🔥 The results are in. The full picture is coming. Stay tuned.
@ekosproject
EKOS has delivered measurable results in its legacy-to-Lakehouse test. But we're not stopping at numbers on a screen.
🎬 We're bringing the entire process to life.
From analyzing thousands of Java files to ainst Oracle, we're preparing a full demo that puts the results front and center.
Code tells the truth. Seeing it happen makes that truth impossible to ignore.
🔥 The results are in. The full picture is coming. Stay tuned.
@ekosproject
🚀 A MAJOR MILESTONE FOR EKOS 🔥
From legacy ERP source code to a reconciled Lakehouse.
What happens when enterprise data pipelines are built on evidence from the source system—not assumptions?
We put EKOS to the test on iDempiere, an open-source Java ERP running on Oracle.
🔥 THE RESULTS:
🔍 4,542 Java files analyzed
📚 1,060 pages of source-system documentation generated
🧪 165 tables tested end-to-end, with every Bronze table reconciled against Oracle
📊 3,664 of 3,667 Silver tests passed
🛠️ 6 real bugs uncovered during testing
⚙️ Zero LLM calls from analysis through Silver
But the real milestone goes beyond the numbers.
EKOS compiled source code, database migrations, Git history, documentation, and the live Oracle catalog into an evidence-backed knowledge ledger.
That evidence informed the pipeline:
→ Trino → Iceberg Bronze
→ dbt-trino Silver
→ Dagster orchestration
Incremental loading is selected based on source-code behavior. Deletes are detected through bucketed key comparisons, with safeguards that hold mass deletions for human approval. Silver tests derive from database constraints and the Java model.
Every pipeline decision carries its evidence.
The full 849-table run remains constrained by available memory, and 195 conflicts still await human review. Gold Mart drafting works; the remaining work is ongoing.
This is a major step toward making legacy modernization more traceable, testable, and grounded in how enterprise systems actually behave.
EKOS: Enterprise knowledge, compiled into engineering evidence.
#EKOS #DataEngineering #Lakehouse #OpenSource4o
@Oracle @OracleCloud @googlecloud @thatjeffsmith @SQLServer @Dell @IBM
2
9
24
285
Ekos Project retweeted
🚀 A MAJOR MILESTONE FOR EKOS 🔥
From legacy ERP source code to a reconciled Lakehouse.
What happens when enterprise data pipelines are built on evidence from the source system—not assumptions?
We put EKOS to the test on iDempiere, an open-source Java ERP running on Oracle.
🔥 THE RESULTS:
🔍 4,542 Java files analyzed
📚 1,060 pages of source-system documentation generated
🧪 165 tables tested end-to-end, with every Bronze table reconciled against Oracle
📊 3,664 of 3,667 Silver tests passed
🛠️ 6 real bugs uncovered during testing
⚙️ Zero LLM calls from analysis through Silver
But the real milestone goes beyond the numbers.
EKOS compiled source code, database migrations, Git history, documentation, and the live Oracle catalog into an evidence-backed knowledge ledger.
That evidence informed the pipeline:
→ Trino → Iceberg Bronze
→ dbt-trino Silver
→ Dagster orchestration
Incremental loading is selected based on source-code behavior. Deletes are detected through bucketed key comparisons, with safeguards that hold mass deletions for human approval. Silver tests derive from database constraints and the Java model.
Every pipeline decision carries its evidence.
The full 849-table run remains constrained by available memory, and 195 conflicts still await human review. Gold Mart drafting works; the remaining work is ongoing.
This is a major step toward making legacy modernization more traceable, testable, and grounded in how enterprise systems actually behave.
EKOS: Enterprise knowledge, compiled into engineering evidence.
#EKOS #DataEngineering #Lakehouse #OpenSource4o
@Oracle @OracleCloud @googlecloud @thatjeffsmith @SQLServer @Dell @IBM
From legacy ERP source code to a reconciled Lakehouse.
What happens when enterprise data pipelines are built on evidence from the source system—not assumptions?
We put EKOS to the test on iDempiere, an open-source Java ERP running on Oracle.
🔥 THE RESULTS:
🔍 4,542 Java files analyzed
📚 1,060 pages of source-system documentation generated
🧪 165 tables tested end-to-end, with every Bronze table reconciled against Oracle
📊 3,664 of 3,667 Silver tests passed
🛠️ 6 real bugs uncovered during testing
⚙️ Zero LLM calls from analysis through Silver
But the real milestone goes beyond the numbers.
EKOS compiled source code, database migrations, Git history, documentation, and the live Oracle catalog into an evidence-backed knowledge ledger.
That evidence informed the pipeline:
→ Trino → Iceberg Bronze
→ dbt-trino Silver
→ Dagster orchestration
Incremental loading is selected based on source-code behavior. Deletes are detected through bucketed key comparisons, with safeguards that hold mass deletions for human approval. Silver tests derive from database constraints and the Java model.
Every pipeline decision carries its evidence.
The full 849-table run remains constrained by available memory, and 195 conflicts still await human review. Gold Mart drafting works; the remaining work is ongoing.
This is a major step toward making legacy modernization more traceable, testable, and grounded in how enterprise systems actually behave.
EKOS: Enterprise knowledge, compiled into engineering evidence.
#EKOS #DataEngineering #Lakehouse #OpenSource4o
@Oracle @OracleCloud @googlecloud @thatjeffsmith @SQLServer @Dell @IBM
You can’t understand a legacy ERP by looking at its database schema alone. 🔍
This week, I put EKOS to the test on iDempiere, an open-source Java ERP supporting Oracle and PostgreSQL.
The challenge? Reconstructing the engineering knowledge hidden across code, SQL, database s logic. ⚙️
🔍 Here are the results:
📊 Java Code Analysis
• 4,542 files parsed, none skipped
• 763 tables and 16,660 columns recovered with business names
• 2,882 table relationships identified in code
• 3,207 SQL statements extracted, with parsing status explicitly recorded
🗄️ Oracle Migration Analysis
• Failed parses reduced from 5,317 to just 9
• 62,251 migration statements analyzed
⚙️ PL/SQL Analysis
• All 61 stored procedures and functions recovered
• 991 of 992 statements parsed
• One remaining gap: a GOTO statement
🧬 Migration History
• 2,154 changesets and 26,060 changes analyzed
• Zero unrecognized changes
• Reported conflicts reduced from 2,372 to just 3, of which 2 are genuine conflicts
💡 Why does this matter?
Legacy modernization requires more than schema extraction. It requires reconstructing business semantics, relationships, SQL logic, procedural code, and how the system evolved over time.
🧠 EKOS turns this scattered engineering knowledge into a structured, evidence-backed representation.
🔗 Every recovered fact links back to its source file and line.
🛠️ No LLM calls. Just rule-based analysis.
📌 This work is currently available on a separate branch.
🚀 Next step: Connect a live Oracle database and compare the running system against the code. Every mismatch becomes a finding traceable to evidence.
#EKOS #DataEngineering #LegacyModernization #ERP #Oracle #OpenSource
@Oracle @OracleCloud @googlecloud @thatjeffsmith
@SQLServer @Dell @IBM
@UberEng @sqldaily @MySQL
@iDempiere @alibaba_cloud
9
11
28
1,271
Ekos Project retweeted
🚀 EKOS UPDATE 🛠️⚙️
We're making EKOS more accessible and easier to evaluate.
🖥️ The latest Web Console update introduces:
🔹 Database source management
🔹 Connection testing directly through the interface
🔹 Clearer visibility into how live database structures compare into the EKOS Ledger
🏢 Why does this matter?
This is an important step toward helping companies, decision-makers, technical teams, and other interested users explore EKOS's real capabilities and evaluate the technology through practical testing.
🔓 What's next?
Free EKOS testing for everyone is on our roadmap. We're working toward making it possible for anyone to experience the technology firsthand, explore its capabilities, and see how it works with real enterprise data.
🎯 Our goal: broader access, greater transparency, and real-world evaluation.
EKOS — Making enterprise knowledge traceable, verifiable, and useful for AI agents.
#EKOS #EnterpriseAI #AIInfrastructure #LegacyModernization
solana:CwubepDFJndzSKFmAMAm9u8Xx3PrizAwSq8hcGimpump @ekosproject
We're making EKOS more accessible and easier to evaluate.
🖥️ The latest Web Console update introduces:
🔹 Database source management
🔹 Connection testing directly through the interface
🔹 Clearer visibility into how live database structures compare into the EKOS Ledger
🏢 Why does this matter?
This is an important step toward helping companies, decision-makers, technical teams, and other interested users explore EKOS's real capabilities and evaluate the technology through practical testing.
🔓 What's next?
Free EKOS testing for everyone is on our roadmap. We're working toward making it possible for anyone to experience the technology firsthand, explore its capabilities, and see how it works with real enterprise data.
🎯 Our goal: broader access, greater transparency, and real-world evaluation.
EKOS — Making enterprise knowledge traceable, verifiable, and useful for AI agents.
#EKOS #EnterpriseAI #AIInfrastructure #LegacyModernization
solana:CwubepDFJndzSKFmAMAm9u8Xx3PrizAwSq8hcGimpump @ekosproject
6
7
37
2,345
🚀 The numbers tell the story. Now, it's time to SHOW the world.
EKOS has delivered measurable results in its legacy-to-Lakehouse test. But we're not stopping at numbers on a screen.
🎬 We're bringing the entire process to life.
From analyzing thousands of Java files to ainst Oracle, we're preparing a full demo that puts the results front and center.
Code tells the truth. Seeing it happen makes that truth impossible to ignore.
🔥 The results are in. The full picture is coming. Stay tuned.
@ekosproject
EKOS has delivered measurable results in its legacy-to-Lakehouse test. But we're not stopping at numbers on a screen.
🎬 We're bringing the entire process to life.
From analyzing thousands of Java files to ainst Oracle, we're preparing a full demo that puts the results front and center.
Code tells the truth. Seeing it happen makes that truth impossible to ignore.
🔥 The results are in. The full picture is coming. Stay tuned.
@ekosproject
🚀 A MAJOR MILESTONE FOR EKOS 🔥
From legacy ERP source code to a reconciled Lakehouse.
What happens when enterprise data pipelines are built on evidence from the source system—not assumptions?
We put EKOS to the test on iDempiere, an open-source Java ERP running on Oracle.
🔥 THE RESULTS:
🔍 4,542 Java files analyzed
📚 1,060 pages of source-system documentation generated
🧪 165 tables tested end-to-end, with every Bronze table reconciled against Oracle
📊 3,664 of 3,667 Silver tests passed
🛠️ 6 real bugs uncovered during testing
⚙️ Zero LLM calls from analysis through Silver
But the real milestone goes beyond the numbers.
EKOS compiled source code, database migrations, Git history, documentation, and the live Oracle catalog into an evidence-backed knowledge ledger.
That evidence informed the pipeline:
→ Trino → Iceberg Bronze
→ dbt-trino Silver
→ Dagster orchestration
Incremental loading is selected based on source-code behavior. Deletes are detected through bucketed key comparisons, with safeguards that hold mass deletions for human approval. Silver tests derive from database constraints and the Java model.
Every pipeline decision carries its evidence.
The full 849-table run remains constrained by available memory, and 195 conflicts still await human review. Gold Mart drafting works; the remaining work is ongoing.
This is a major step toward making legacy modernization more traceable, testable, and grounded in how enterprise systems actually behave.
EKOS: Enterprise knowledge, compiled into engineering evidence.
#EKOS #DataEngineering #Lakehouse #OpenSource4o
@Oracle @OracleCloud @googlecloud @thatjeffsmith @SQLServer @Dell @IBM
2
9
24
285
Ekos Project retweeted
The real opportunity goes far beyond a single test. 🚀
Enterprises spend enormous resources dealing with legacy systems, hidden bugs, and fragmented knowledge. Solving these problems creates real business value.
@ekosproject is turning enterprise knowledge into verifiable, — with growing potential for real-world adoption.⚙️
We’re also open to collaborating with individuals and companies who recognize this potential and want to help move it forward.🤝👷♂️
Real technology. Real problems. Long-term potential.💎
@sama @GavinSBaker @cz_binance
@Cognizant @connor_mc_d @CompleteSkeptic @elonmusk @ericzakariasson @finkd @paulg @rauchg @saranormous @JensenHuang
@hwchase17 @romainhuet @jeff_weinstein @OfficialLoganK
Enterprises spend enormous resources dealing with legacy systems, hidden bugs, and fragmented knowledge. Solving these problems creates real business value.
@ekosproject is turning enterprise knowledge into verifiable, — with growing potential for real-world adoption.⚙️
We’re also open to collaborating with individuals and companies who recognize this potential and want to help move it forward.🤝👷♂️
Real technology. Real problems. Long-term potential.💎
@sama @GavinSBaker @cz_binance
@Cognizant @connor_mc_d @CompleteSkeptic @elonmusk @ericzakariasson @finkd @paulg @rauchg @saranormous @JensenHuang
@hwchase17 @romainhuet @jeff_weinstein @OfficialLoganK
🚀 A MAJOR MILESTONE FOR EKOS 🔥
From legacy ERP source code to a reconciled Lakehouse.
What happens when enterprise data pipelines are built on evidence from the source system—not assumptions?
We put EKOS to the test on iDempiere, an open-source Java ERP running on Oracle.
🔥 THE RESULTS:
🔍 4,542 Java files analyzed
📚 1,060 pages of source-system documentation generated
🧪 165 tables tested end-to-end, with every Bronze table reconciled against Oracle
📊 3,664 of 3,667 Silver tests passed
🛠️ 6 real bugs uncovered during testing
⚙️ Zero LLM calls from analysis through Silver
But the real milestone goes beyond the numbers.
EKOS compiled source code, database migrations, Git history, documentation, and the live Oracle catalog into an evidence-backed knowledge ledger.
That evidence informed the pipeline:
→ Trino → Iceberg Bronze
→ dbt-trino Silver
→ Dagster orchestration
Incremental loading is selected based on source-code behavior. Deletes are detected through bucketed key comparisons, with safeguards that hold mass deletions for human approval. Silver tests derive from database constraints and the Java model.
Every pipeline decision carries its evidence.
The full 849-table run remains constrained by available memory, and 195 conflicts still await human review. Gold Mart drafting works; the remaining work is ongoing.
This is a major step toward making legacy modernization more traceable, testable, and grounded in how enterprise systems actually behave.
EKOS: Enterprise knowledge, compiled into engineering evidence.
#EKOS #DataEngineering #Lakehouse #OpenSource4o
@Oracle @OracleCloud @googlecloud @thatjeffsmith @SQLServer @Dell @IBM
4
10
29
480
🚀 A MAJOR MILESTONE FOR EKOS 🔥
From legacy ERP source code to a reconciled Lakehouse.
What happens when enterprise data pipelines are built on evidence from the source system—not assumptions?
We put EKOS to the test on iDempiere, an open-source Java ERP running on Oracle.
🔥 THE RESULTS:
🔍 4,542 Java files analyzed
📚 1,060 pages of source-system documentation generated
🧪 165 tables tested end-to-end, with every Bronze table reconciled against Oracle
📊 3,664 of 3,667 Silver tests passed
🛠️ 6 real bugs uncovered during testing
⚙️ Zero LLM calls from analysis through Silver
But the real milestone goes beyond the numbers.
EKOS compiled source code, database migrations, Git history, documentation, and the live Oracle catalog into an evidence-backed knowledge ledger.
That evidence informed the pipeline:
→ Trino → Iceberg Bronze
→ dbt-trino Silver
→ Dagster orchestration
Incremental loading is selected based on source-code behavior. Deletes are detected through bucketed key comparisons, with safeguards that hold mass deletions for human approval. Silver tests derive from database constraints and the Java model.
Every pipeline decision carries its evidence.
The full 849-table run remains constrained by available memory, and 195 conflicts still await human review. Gold Mart drafting works; the remaining work is ongoing.
This is a major step toward making legacy modernization more traceable, testable, and grounded in how enterprise systems actually behave.
EKOS: Enterprise knowledge, compiled into engineering evidence.
#EKOS #DataEngineering #Lakehouse #OpenSource4o
@Oracle @OracleCloud @googlecloud @thatjeffsmith @SQLServer @Dell @IBM
From legacy ERP source code to a reconciled Lakehouse.
What happens when enterprise data pipelines are built on evidence from the source system—not assumptions?
We put EKOS to the test on iDempiere, an open-source Java ERP running on Oracle.
🔥 THE RESULTS:
🔍 4,542 Java files analyzed
📚 1,060 pages of source-system documentation generated
🧪 165 tables tested end-to-end, with every Bronze table reconciled against Oracle
📊 3,664 of 3,667 Silver tests passed
🛠️ 6 real bugs uncovered during testing
⚙️ Zero LLM calls from analysis through Silver
But the real milestone goes beyond the numbers.
EKOS compiled source code, database migrations, Git history, documentation, and the live Oracle catalog into an evidence-backed knowledge ledger.
That evidence informed the pipeline:
→ Trino → Iceberg Bronze
→ dbt-trino Silver
→ Dagster orchestration
Incremental loading is selected based on source-code behavior. Deletes are detected through bucketed key comparisons, with safeguards that hold mass deletions for human approval. Silver tests derive from database constraints and the Java model.
Every pipeline decision carries its evidence.
The full 849-table run remains constrained by available memory, and 195 conflicts still await human review. Gold Mart drafting works; the remaining work is ongoing.
This is a major step toward making legacy modernization more traceable, testable, and grounded in how enterprise systems actually behave.
EKOS: Enterprise knowledge, compiled into engineering evidence.
#EKOS #DataEngineering #Lakehouse #OpenSource4o
@Oracle @OracleCloud @googlecloud @thatjeffsmith @SQLServer @Dell @IBM
You can’t understand a legacy ERP by looking at its database schema alone. 🔍
This week, I put EKOS to the test on iDempiere, an open-source Java ERP supporting Oracle and PostgreSQL.
The challenge? Reconstructing the engineering knowledge hidden across code, SQL, database s logic. ⚙️
🔍 Here are the results:
📊 Java Code Analysis
• 4,542 files parsed, none skipped
• 763 tables and 16,660 columns recovered with business names
• 2,882 table relationships identified in code
• 3,207 SQL statements extracted, with parsing status explicitly recorded
🗄️ Oracle Migration Analysis
• Failed parses reduced from 5,317 to just 9
• 62,251 migration statements analyzed
⚙️ PL/SQL Analysis
• All 61 stored procedures and functions recovered
• 991 of 992 statements parsed
• One remaining gap: a GOTO statement
🧬 Migration History
• 2,154 changesets and 26,060 changes analyzed
• Zero unrecognized changes
• Reported conflicts reduced from 2,372 to just 3, of which 2 are genuine conflicts
💡 Why does this matter?
Legacy modernization requires more than schema extraction. It requires reconstructing business semantics, relationships, SQL logic, procedural code, and how the system evolved over time.
🧠 EKOS turns this scattered engineering knowledge into a structured, evidence-backed representation.
🔗 Every recovered fact links back to its source file and line.
🛠️ No LLM calls. Just rule-based analysis.
📌 This work is currently available on a separate branch.
🚀 Next step: Connect a live Oracle database and compare the running system against the code. Every mismatch becomes a finding traceable to evidence.
#EKOS #DataEngineering #LegacyModernization #ERP #Oracle #OpenSource
@Oracle @OracleCloud @googlecloud @thatjeffsmith
@SQLServer @Dell @IBM
@UberEng @sqldaily @MySQL
@iDempiere @alibaba_cloud
9
11
28
1,271
This is just the beginning. Imagine what becomes possible when legacy systems can be understood, tested, and modernized through verifiable evidence. 🛠️
EKOS is building toward a future where enterprise knowledge becomes a foundation for smarter, more reliable AI and data ential is enormous.
@ekosproject solana:CwubepDFJndzSKFmAMAm9u8Xx3PrizAwSq8hcGimpump
EKOS is building toward a future where enterprise knowledge becomes a foundation for smarter, more reliable AI and data ential is enormous.
@ekosproject solana:CwubepDFJndzSKFmAMAm9u8Xx3PrizAwSq8hcGimpump
1
0
3
24
Real progress isn’t just about building something new. It’s about testing it against real systems, finding hidden bugs, and verifying that the results match the source.
EKOS analyzed 4,542 Java files, tested 165 tables against Oracle, and uncovered 6 real bugs.
That’s how you turn technical claims into measurable evidence.
medium.com/@alexeyban/fro…
EKOS analyzed 4,542 Java files, tested 165 tables against Oracle, and uncovered 6 real bugs.
That’s how you turn technical claims into measurable evidence.
medium.com/@alexeyban/fro…
0
1
6
130