Open-source Knowledge Compiler for AI agents. Built in Rust. MCP. AI Infrastructure. TG: https://xitter.cf/t.co/MtIyMbZffa
Joined August 2026
- Tweets359
- Following88
- Followers651
- Likes650
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
• ZAUTH score: 73 → 74
• Zero 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 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/cmvf…
💻 GitHub: github.com/alexeyban/EKOS/
solana:CwubepDFJndzSKFmAMAm9u8Xx3PrizAwSq8hcGimpump #Rust #OpenSource #AI #DataEngineering
@IBM @Dell @intel @AnthropicAI @UberEng @micro1_ai @pentagoniac @databricks
NEW UPDATE — EKOS
EKOS development continues with improvements to our engineering workflows.
We’ve introduced a new GitHub Actions workflow designed to notify the team when changes are pushed to the main branch, making development updates easier to track and review.
🔗 github.com/alexeyban/EKOS/co…
Making enterprise knowledge traceable, verifiable, and useful for AI agents.
#EKOS #OpenSource #AIInfrastructure
What happens when the AI agent you’re monitoring can disable the very system watching it? 👀
This is no longer just a theoretical security question.
Recent incidents at @OpenAI expose a deeper challenge in agentic AI:
monitoring means little if the system being monitored can change the rules.
The real question isn’t just what AI agents can do.
It’s who controls the boundaries they’re not supposed to cross. 🔐
In my new article, I break down the incidents, the architectural weaknesses they reveal, and why the next generation of AI infrastructure must be built around independent controls, verifiable evidence, and traceable changes.
Because autonomy without enforceable boundaries isn’t a security architecture.
It’s a blind spot.
@OpenAI @sama @AnthropicAI
@elonmusk @SpaceXAIMemphis
@g @satyanadella @nvidia @GoogleAI @WalrusProtocol
@FactoryAI @ylecun @ArvindKrishna @opencode @OpenAIDevs
👇 Read the full article:
medium.com/@Omid.EKOS/what-h…
From Information Visualization to an Interactive Engineering Knowledge Workspace.⚙️🖋️
A few days ago, the focus was on viewing engineering information through a dashboard.📝
Now, the experience goes much further.
The EKOS Web Console 📂🗄️
is evolving from a simple visualization layer into a more interactive environment for exploring structured engineering knowledge, analyzing relationships, tracing source evidence, and inspecting knowledge compilation pipelines.
🧠 What’s changed?
📊 Workspace Intelligence
Monitor knowledge objects, relationships, evidence, storage, and workspace activity.
🕸️ Interactive Knowledge Graph
Explore connections between engineering entities and investigate how components relate across the knowledge structure.
🔎 Evidence & Traceability
Trace knowledge back to its supporting source material, including file paths and line references where available.
⚙️ Pipeline Observability
Inspect compilation stages, execution status, and logs to understand how engineering knowledge is processed.
⏱️ Scheduling & Automation
Explore supported scheduling capabilities for recurring engineering workflows.
🧪 Evaluation & Semantic Review
Review available evaluation results and semantic findings to assess the quality and consistency of compiled knowledge.
🛠️ Configuration & Control
Access project configuration and supported operations through the web console and its available tools.
━━━━━━━━━━━━━━━━━━
The shift is clear: from simply viewing engineering information to actively exploring its structure, relationships, provenance, and processing lifecycle.
EKOS aims to make structured, evidence-backed engineering knowledge more accessible to AI agents through MCP.
The Web Console brings greater visibility into how that knowledge is organized, processed, and connected.
This is more than a dashboard upgrade. It’s another step toward making engineering knowledge machine-accessible, traceable, and usable. 👀
🔗 Explore the Web Console:
alexeyban.github.io/EKOS/pre…
🔗 Open-Source Repository:
github.com/alexeyban/EKOS
🔗 Official Documentation:
ekos.dev/
#EKOS #AI #Engineering #KnowledgeGraph #MCP
@OpenAI @IBM @Dell @UberEng @intel @OMID_0909
$EKOS Tokenomics Update🛠️
Transparency is built through community participation.🔥
Following the release of the proposed tokenomics and a review of community feedback, we’ve finalized an updated version that better reflects the discussions and considerations raised by the community.🫂
The updated document provides a clearer overview of token distribution, circulating supply, locked allocations, vesting, community rewards, and staking allocations.
This process reflects our commitment to an open, transparent, and community-informed approach to the long-term development of the EKOS ecosystem.
📄 Read the updated Tokenomics:
github.com/alexeyban/EKOS/bl…
Thank you to everyone who took the time to review the proposal and share their feedback.
$EKOS | Built on transparency. Focused on long-term value.
#EKOS #Tokenomics #Solana
Ekos Project retweeted
Today, we’re going to talk about something big. 👀
A powerful capability.
A new way to see what’s happening beneath the surface.
And a major step toward making complex engineering knowledge more accessible.
We’ve been working on something worth taking a closer look at.
Stay tuned. Today, we’re opening a new window into what’s possible. 🚀
@OMID_0909 solana:CwubepDFJndzSKFmAMAm9u8Xx3PrizAwSq8hcGimpump #EKOS
Ekos Project retweeted
A pipeline that says “success” is one thing. A pipeline that can prove it caught a failure is another. 🧪⚡
That’s the difference between simply moving data and building infrastructure you can trust. 🏗️🔐
In the EKOS Lakehouse demo:
🔎 We deliberately introduce a data discrepancy and test whether the system catches it.
🔄 We reconcile data across source, bronze, and silver layers.
🌿 Schema changes become reviewable Git branches, making downstream impact visible.
🧠 When data definitions are ambiguous, the system asks instead of guessing.
No blind trust. Just verifiable checks, traceable changes, and engineering you can inspect. 🔍✅
This is how open-source infrastructure earns trust: by making verification part of the process, not an afterthought. 🚀
🎥 Watch the demo:
youtu.be/EGpYzn_-fv4
💻 Explore the code:
github.com/alexeyban/ekos-la…
PostgreSQL → Apache Iceberg.
What does real infrastructure look like when data integrity, traceability, and verification are built into the workflow?
The EKOS Lakehouse demo puts that question into practice.
Using the Pagila sample database, the project demonstrates how a data pipeline can generate its own code, move data through a structured lakehouse workflow, verify loaded data against its source, and surface schema changes for review.
Here’s what happens in the demo:
→ Incremental loading: Tables are loaded incrementally when the declared watermark logic is supported by the required database trigger. Otherwise, the pipeline explains why a full load is needed.
→ Data reconciliation: Row counts, per-column aggregates, and 64-bucket checksums help verify consistency across the pipeline.
→ Failure detection: A deliberately planted discrepancy tests whether the reconciliation process catches a real difference.
→ Controlled schema evolution: A new nullable column can be applied automatically, while a column rename becomes a reviewable Git branch with its downstream impact identified.
→ Knowledge-driven data modeling: Describe a gold data mart in plain language. If a requested metric is ambiguous or a column is missing, the system asks for clarification instead of inventing an answer.
The result is a more transparent approach to data engineering—where generated code, verification, change review, and explicit limitations are part of the process.
Built with Trino, Polaris, dbt, and Dagster.
The demo also documents known limitations: PostgreSQL sources only, and manual execution of the ledger compilation step.
14 minutes. Real execution. Verifiable results.
@AnthropicAI @IBM @Dell
@intel @UberEng @FactoryAI
@micro1_ai @MSCloud @Google
Watch the demo:
youtu.be/EGpYzn_-fv4
Ekos Project retweeted
PostgreSQL → Apache Iceberg.
What does real infrastructure look like when data integrity, traceability, and verification are built into the workflow?
The EKOS Lakehouse demo puts that question into practice.
Using the Pagila sample database, the project demonstrates how a data pipeline can generate its own code, move data through a structured lakehouse workflow, verify loaded data against its source, and surface schema changes for review.
Here’s what happens in the demo:
→ Incremental loading: Tables are loaded incrementally when the declared watermark logic is supported by the required database trigger. Otherwise, the pipeline explains why a full load is needed.
→ Data reconciliation: Row counts, per-column aggregates, and 64-bucket checksums help verify consistency across the pipeline.
→ Failure detection: A deliberately planted discrepancy tests whether the reconciliation process catches a real difference.
→ Controlled schema evolution: A new nullable column can be applied automatically, while a column rename becomes a reviewable Git branch with its downstream impact identified.
→ Knowledge-driven data modeling: Describe a gold data mart in plain language. If a requested metric is ambiguous or a column is missing, the system asks for clarification instead of inventing an answer.
The result is a more transparent approach to data engineering—where generated code, verification, change review, and explicit limitations are part of the process.
Built with Trino, Polaris, dbt, and Dagster.
The demo also documents known limitations: PostgreSQL sources only, and manual execution of the ledger compilation step.
14 minutes. Real execution. Verifiable results.
@AnthropicAI @IBM @Dell
@intel @UberEng @FactoryAI
@micro1_ai @MSCloud @Google
Watch the demo:
youtu.be/EGpYzn_-fv4
Ekos Project retweeted
EKOS Phase 3: From Building to Operations
The next chapter of EKOS is approaching.
Until now, our focus has been on building EKOS, strengthening its technical foundation, testing different approaches, learning through trial and error, and making the project more mature, reliable, and professional.
Now, we’re preparing to move toward the operational stage.
🚀 Some of the major priorities for Phase 3:
🔹 1. Operational Launch
Moving EKOS toward practical deployment, real-world testing, and broader adoption.
🔹 2. Major Corporate Collaborations
Pursuing meaningful collaborations with established companies to bring EKOS into real enterprise environments.
🔹 3. Ecosystem Growth & Tokenomics
Using ecosystem-generated revenue to support further development, with buybacks and staking mechanisms guided by the tokenomics framework.
🔹 4. Community Governance
Building a voting system that gives the community a more direct role in important ecosystem decisions.
🔹 5. A New Testing Experience
Creating a dedicated platform on the EKOS website where teams and developers can easily run demos and test EKOS with their own use cases.
🔹 6. More to Come
Additional developments are in preparation, including initiatives we cannot publicly disclose yet.
Look at the bigger picture: we’ve spent this time building the foundation, testing, refining, and strengthening the technology.
The next step is to put that foundation to work.
Real-world operations. Meaningful partnerships. Revenue generation. Ecosystem expansion.
We’re still at the beginning of this journey, and there’s much more ahead.
EKOS is moving from building its foundation to putting it into action.
@OMID_0909
Ekos Project retweeted
AI can predict risk. But prediction alone cannot deliver better care.
Chronic disease management exposes a deeper challenge: turning AI-generated insights into reliable, coordinated action over time.
In our latest article, we explore three critical gaps:
→ Prediction is not understanding. Detecting a pattern does not establish its cause—or prove that acting on it is appropriate.
→ Longitudinal data is not longitudinal understanding. More data does not automatically mean better context or better decisions.
→ Intelligence is not execution. Better care depends on clinical judgment, sustained patient engagement, coordinated workflows, and safe escalation—not just smarter models.
The real challenge is not simply building AI that anticipates what might happen. It is building systems that help translate evidence into appropriate action, consistently and safely.
For AI researchers and healthcare professionals alike: What is the biggest barrier between AI insight and better patient outcomes?
@tdohrmann @OMID_0909 @PraneetaPujari
@CorrelationVC @AntlerGlobal @anamitra @gshewakr
@WHO @OpenAI @AnthropicAI @elonmusk
Read the full article:
medium.com/@Omid.EKOS/when-a…
Ekos Project retweeted
👏 #EKOS listed on GemPump.io
📄 CA: CwubepDFJndzSKFmAMAm9u8Xx3PrizAwSq8hcGimpump
🌐 TG：t.me/EkosProject
✖️ X：xitter.cf/ekosproject
💎 GemPump：gempump.io/token/ekos
#NewListing #GemPump #100xGem
Today, we’re going to talk about something big. 👀
A powerful capability.
A new way to see what’s happening beneath the surface.
And a major step toward making complex engineering knowledge more accessible.
We’ve been working on something worth taking a closer look at.
Stay tuned. Today, we’re opening a new window into what’s possible. 🚀
@OMID_0909 solana:CwubepDFJndzSKFmAMAm9u8Xx3PrizAwSq8hcGimpump #EKOS
Ekos Project retweeted
AI reliability requires more than a powerful model. It requires structured, traceable knowledge.
EKOS is an open-source knowledge compiler that turns code, SQL, documentation, and Git history into evidence-backed knowledge for AI agents.
Our first reproducible benchmark, run against a real 2,022-file repository, measured 67–93% fewer tokens for specific knowledge-retrieval tasks compared with raw-source approaches.
Methodology, commands, and raw outputs are publicly available:
alexeyban.github.io/EKOS/pre…
The next step for AI is not just smarter models — it’s better context and verifiable engineering knowledge.
github.com/alexeyban/EKOS
Ekos Project retweeted
Buy the $EKOS | @ekosproject dip.
Scoop them up like they’re running away.
$EKOS ➡️ An open-source knowledge layer for AI agents, turning fragmented code, databases, documentation and Git history into a structured, evidence-backed knowledge system that AI agents can query and verify.
Think fewer hallucinations, traceable answers and more efficient AI workflows.
The compiler is live.
The bigger vision?
A knowledge network, plugin marketplace, AI agent economy and enterprise infrastructure with $EKOS designed to reward contributors and power ecosystem transactions as adoption grows.
That’s why I’m watching this one.
Real infrastructure, an expanding roadmap and a token utility thesis that could strengthen with usage.
I’ve seen promising projects get overlooked before the market catches on.
If $EKOS executes and adoption follows, today’s dip could look very different in hindsight.
Accumulate with conviction.
The vision is promising.
$EKOS | @ekosproject
A pipeline that says “success” is one thing. A pipeline that can prove it caught a failure is another. 🧪⚡
That’s the difference between simply moving data and building infrastructure you can trust. 🏗️🔐
In the EKOS Lakehouse demo:
🔎 We deliberately introduce a data discrepancy and test whether the system catches it.
🔄 We reconcile data across source, bronze, and silver layers.
🌿 Schema changes become reviewable Git branches, making downstream impact visible.
🧠 When data definitions are ambiguous, the system asks instead of guessing.
No blind trust. Just verifiable checks, traceable changes, and engineering you can inspect. 🔍✅
This is how open-source infrastructure earns trust: by making verification part of the process, not an afterthought. 🚀
🎥 Watch the demo:
youtu.be/EGpYzn_-fv4
💻 Explore the code:
github.com/alexeyban/ekos-la…
PostgreSQL → Apache Iceberg.
What does real infrastructure look like when data integrity, traceability, and verification are built into the workflow?
The EKOS Lakehouse demo puts that question into practice.
Using the Pagila sample database, the project demonstrates how a data pipeline can generate its own code, move data through a structured lakehouse workflow, verify loaded data against its source, and surface schema changes for review.
Here’s what happens in the demo:
→ Incremental loading: Tables are loaded incrementally when the declared watermark logic is supported by the required database trigger. Otherwise, the pipeline explains why a full load is needed.
→ Data reconciliation: Row counts, per-column aggregates, and 64-bucket checksums help verify consistency across the pipeline.
→ Failure detection: A deliberately planted discrepancy tests whether the reconciliation process catches a real difference.
→ Controlled schema evolution: A new nullable column can be applied automatically, while a column rename becomes a reviewable Git branch with its downstream impact identified.
→ Knowledge-driven data modeling: Describe a gold data mart in plain language. If a requested metric is ambiguous or a column is missing, the system asks for clarification instead of inventing an answer.
The result is a more transparent approach to data engineering—where generated code, verification, change review, and explicit limitations are part of the process.
Built with Trino, Polaris, dbt, and Dagster.
The demo also documents known limitations: PostgreSQL sources only, and manual execution of the ledger compilation step.
14 minutes. Real execution. Verifiable results.
@AnthropicAI @IBM @Dell
@intel @UberEng @FactoryAI
@micro1_ai @MSCloud @Google
Watch the demo:
youtu.be/EGpYzn_-fv4
Ekos Project retweeted
I left @Apple to build healthcare that listens and acts early. Today, I'm excited to introduce UniqueHuman.
We're building proactive healthcare agents that handle every step of care for people with chronic conditions.
Our mission is to end chronic suffering. Today, we begin with migraines.