AI Engineer • Full-Stack Builder • Workflow Automator
Designing intelligence that reasons, acts, and scales

I build agentic software systems where LLMs don’t just generate text — they plan, use tools, execute workflows, and validate results.

My focus areas:

• Multi-step AI agents with memory & orchestration
• RAG pipelines over real-world datasets
• Streaming + async backends for long-running agent tasks
• Automation-first architectures with queues and workers

I care about clean modular design, observability, and production-ready AI.

AI Engineering
LangChain · LangGraph · OpenAI API · RAG · Tool Calling · Vector Stores

Backend
FastAPI · Python · PostgreSQL · SQLite · REST · Webhooks

Frontend
React · TypeScript · Tailwind · Streamlit

Data & Orchestration
Redis · Celery · Async workers · Event streaming

Infra & DevOps
Docker · GitHub Actions · API-first architecture

Multi-agent system for Argo float oceanographic data with end-to-end scientific workflows.

Agent Graph:
query_understanding → data_retrieval → analysis → validation

Capabilities:

• Natural language parsing → location, depth, time range, variables
• ERDDAP data ingestion + SQLite analytical layer
• Validation engine with:
  • confidence scoring
  • time-window consistency checks
  • outlier-rate detection
• Streamlit control panel with:
  • live execution progress events
  • diagnostics cards
  • Plotly visualizations (trends, depth profiles, geo maps, correlations)

APIs:

• /v1/query → structured multi-agent execution
• /v1/stream/{conversation_id} → real-time progress streaming

➡️ Agent orchestration for scientific data pipelines

Production-style lead generation backend with enrichment, scoring, and campaign workflows.

Architecture:
FastAPI · LangChain · OpenAI · PostgreSQL · Redis · Celery · Docker

Features:

• Multi-source lead extraction
  • LinkedIn
  • business directories
  • job boards
• AI enrichment:
  • company intelligence
  • pain-point detection
  • lead scoring
• Campaign orchestration engine
• Persistent SQLite lead cache (deduplication + query reuse)
• Strict real-data mode (no mock fallbacks)
• Async workers for scalable processing
• Interactive FastAPI docs for pipeline testing

➡️ AI + data pipelines + distributed task execution

✔ Design for tool use, not just text output
✔ Prefer graphs over chains for complex reasoning
✔ Separate LLM logic from business logic
✔ Build API-first, agent-ready backends
✔ Optimize for real users and real data

• Long-running agents & background execution models
• Memory architectures (vector + relational + episodic)
• Evaluation & observability for LLM systems
• MLOps for agent deployment

I’m open to:

• AI agent infrastructure projects
• Full-stack AI products
• Research → production LLM systems
• Hackathons & experimental builds

LinkedIn • GitHub • Twitter • Instagram