AI Systems and Web3 Infrastructure Engineer focused on deterministic AI workflows, multi-chain blockchain infrastructure, distributed systems, retrieval-augmented systems, developer tooling, and reliability-focused automation.

I work across AI systems and blockchain infrastructure, with experience across Ethereum, Solana, Sui, Stellar, EVM systems, smart contracts, RPC workflows, indexing, transaction processing, and Web3 developer tooling.

My engineering style is agent-assisted but evidence-first: I use AI-assisted development tools to move faster, while keeping human ownership over debugging, validation, tests, code review, and correctness.

I use AI agents as engineering accelerators, not replacements for engineering judgment.

Part 2 of my CI debugging series.

I wrote about why AI debugging tools should not just produce confident answers. They need evidence-backed diagnoses, structured artifacts, validation gates, observable decisions, and deterministic guardrails.

Part 1 of the series.

I wrote about the design behind ci-rootcause, a GitHub App-first CI failure analysis system that turns failed GitHub Actions runs into evidence-backed RCA comments and structured artifacts.

The article covers deterministic RCA pipelines, diff-aware root-cause ranking, optional LLM-assisted fix proposals, guarded fix PRs, evals, and reproducibility.

Protocol assurance workspace for smart contracts.

ProofBoard helps teams turn protocol intent into executable invariants, verification evidence, and unresolved assumption debt. It is designed around a simple principle: AI proposes, humans approve, and tools produce evidence.

• Protocol intent mapping
• Intent board workflows
• Invariant discovery and approval
• Assumption debt tracking
• Verification evidence ledgers
• Foundry invariant harness export
• Foundry result parsing
• Audit packet export
• ERC4626-style vault assurance
• Human-approved, AI-assisted protocol review

Improve smart contract assurance by connecting AI-assisted protocol analysis with human approval, executable invariants, fuzzing evidence, structured verification artifacts, and unresolved risk tracking.

Deterministic AI-assisted CI failure investigation platform for GitHub Actions.

• Workflow execution reconstruction
• Root-cause ranking with confidence scoring
• Structured RCA artifact generation
• Evidence-backed failure analysis
• AI-assisted fix generation with guardrails
• GitHub App-first architecture
• Multi-model orchestration support

Python, GitHub Actions, OpenAI, Anthropic, Gemini, Ollama, structured pipelines, observability workflows

I contribute to real production systems across AI infrastructure, blockchain infrastructure, and distributed systems.

Current focus areas:

• AI orchestration and agent workflows
• Ethereum infrastructure and Web3 tooling
• Multi-chain blockchain infrastructure across Ethereum, Solana, Sui, and Stellar
• Rust-based distributed systems
• CI/CD and developer infrastructure
• Reliability, correctness, and failure analysis
• Deterministic, test-backed engineering workflows

Approach:

• work on real issues in active repositories
• prioritize maintainable fixes over surface-level changes
• use AI-assisted development for speed, context navigation, and iteration
• keep ownership over validation, tests, debugging, and final implementation quality

I am actively expanding and contributing across Web3 infrastructure ecosystems, especially Ethereum and Rust-based infrastructure today, while also working with Solana, Sui, Stellar, smart contract systems, transaction workflows, and multi-chain backend integrations.

• [Open] #3996 fix(provider): poll receipts while waiting for confirmations in alloy-rs/alloy
• [Open] #24337 fix(download): avoid checksum scan during resume startup in paradigmxyz/reth
• [Open] #8957 Fix 8935 match except dev in qdrant/qdrant
• [Open] #6535 fix(ethereum): handle trace_filter traces missing result.output via c… in graphprotocol/graph-node
• [Open] #2331 fix(langgraph): handle null thread checkpoint in RemoteGraph.getState in langchain-ai/langgraphjs
• [Open] #5461 fix(converter): fall back on invalid JSON-like partial matches in crewAIInc/crewAI
• [Open] #21386 fix(azureaisearch): preserve falsy metadata values in index mapping in run-llama/llama_index
• [Open] #21336 fix(elasticsearch): split sync and async store paths in run-llama/llama_index
• [Merged] #39169 fix(gdn): Align prefill warmup with real prefill path in vllm-project/vllm
• [Open] #10 Fix #9: update guest code for current nssa_core program API in logos-co/logos-lez-rln

• Multi-agent systems
• AI evals
• Structured outputs
• LLM orchestration
• Retrieval workflows
• Context management
• AI observability
• Agent-assisted engineering workflows

• Ethereum and EVM infrastructure
• Solana program and transaction workflows
• Sui / Move-based smart contract systems
• Stellar smart contract and payment systems
• Rust-based Web3 tooling
• RPC and provider systems
• Transaction lifecycle handling
• Node infrastructure
• Indexing and trace processing
• Smart contract systems
• Gas, performance, and reliability optimization

• Distributed systems
• Event-driven architectures
• CI/CD systems
• Reliability engineering
• Async execution systems
• Failure isolation
• Structured tracing
• Regression testing

• Rust
• Python
• TypeScript
• Solidity
• Move

• Ethereum
• Solana
• Sui
• Stellar
• Alloy
• Reth
• Graph Node
• Qdrant
• LangGraph
• CrewAI
• LangChain
• Ollama
• OpenAI
• Anthropic
• GitHub Actions

• Deterministic over opaque automation
• Evidence-backed engineering over confident guesses
• Reliability before complexity
• Tests and validation before claims of completion
• Structured outputs over unbounded generation
• Reproducible execution workflows
• Observability-first system design
• Human-owned, AI-assisted development
• Guardrailed automation for high-impact systems

• GitHub
• LinkedIn
• Medium