Prompt
Guide existing capabilities with structured instructions & formats.
- Prompt
- RAG
- Fine-Tuning
- Agents
I adapt foundation models into production—fast and compliant.
I build robust applications on top of foundation models — starting with the simplest effective lever and scaling only when needed: Prompt → RAG → Fine-Tuning → Agents. My edge: production pragmatism and governance (HIPAA, GDPR, EU AI Act).
Shipped with a pragmatic ladder—Prompt → RAG → Fine-Tuning → Agents.
Guide existing capabilities with structured instructions & formats.
Inject private, current, or domain knowledge at query time.
Teach new behaviors & formats when base model behavior falls short.
Coordinate multi-step tasks, tools, and roles with durable state.
AI-Driven YouTube Growth
SaaS for creators: pulls YouTube data, generates AI SEO suggestions, and plans content strategy.
AI-Powered Regulatory Intelligence
End-to-end platform ingesting healthcare regulations; FTS + RAG for fast, cited answers and risk flags.
Pre-Collection Legal Risk Analysis
Developer-first tool scoring ToS/robots.txt risk pre-scrape; Scrapy/Playwright pipeline + lightweight ML.
Automated Personal Data Removal
Hybrid automation (Celery + Playwright-stealth) with “Guided DIY” for hard targets; Mailpit for verifications.
Interactive AI Engineering Course
Docs-as-code educational platform (MkDocs + Mermaid) shipped via CI/CD to Vercel; open-source curriculum.
Secure Context via Tools
Model Context Protocol server exposes tools/resources; client LLM accesses approved data and vectors.
Autonomous ML First-Pass
ReAct agent selects tools to analyze uploaded CSVs; produces plots and summaries automatically.
Personal Data Exposure & Removal Backend
FastAPI + Celery + Crawl4AI + Playwright; AI validation via OpenRouter; full JWT auth.
Compare RAG Strategies
Streamlit app comparing retrievers (Vector, Multi-Query, Self-Query, Parent Doc); retriever factory pattern.
AI-Powered Contract Review
Structured extraction + risk scoring with LangChain + IBM watsonx; Gradio front-end.
Live Repo Insights & Q&A
Pulls README/metadata via GitHub API; builds a query engine for icebreakers and Q&A.
AI-Assisted Job Search CRM
Next.js + FastAPI + spaCy for JD analysis; PostgreSQL + async API; Dockerized services incl. Redis cache.
Advanced Retrieval Techniques
Script-based lab for hybrid and hierarchical retrieval; BM25 + embeddings, fusion modes, auto-merge.
Reason → Act → Observe
Tool-using agent with web search and custom Python tools; typed state via StateGraph.
Multilingual RAG for Special Education
Secure, multilingual RAG assistant to guide parents through special education processes; containerized stack.
Ask Questions over Your PDFs
Upload a PDF and query it with a vanilla RAG pipeline; Gradio UI and Chroma vector index.
AI Paralegal for Immigration
Private RAG assistant for immigration research; ingest pipelines (PDF/DOCX/web), bias testing, audit logs.
Summarize from a Public URL
End-to-end local RAG for URL pages; quick Gradio interface for demos.
Full-Stack AI Compliance Suite
RAG chatbot + live site scanner + fine-tuned model; Next.js + FastAPI architecture and deployment.
Search → Verify → Revise
Combines Reflection with tool use; structured outputs via Pydantic and robust parsing.
Generate → Critique → Improve
Autonomous reflection loop using LangGraph; structured self-critique and refinement.
Local-first Autonomous Email Agent
Containerized, privacy-first agent that triages Gmail locally; action routing, logging, and observability.
Interactive OSS Catalog
Next.js + D3.js data viz over Supabase; RLS for secure auth; automated data enrichment.
Semantic HR Assistant
Persistent Chroma in client-server mode; advanced metadata filters for talent search.
Telegram RAG with Source Verification
Ingests Telegram chats, builds index, answers with citations; async API integration via Telethon.
Local Multi-Modal Knowledge Base
Local pipeline: Whisper → embeddings → FAISS; query audio/video/text privately via Streamlit UI.
Semantic Search from Scratch
FAISS-based semantic search engine; embeddings, fast indexing, similarity search.
10+ years building privacy-first GenAI systems. Recent highlights below.
Design and deploy privacy-first GenAI platforms across healthcare, legal, and education; specialize in RAG, agents, and compliant cloud delivery.
Shipped secure, high-traffic web apps with React/Node, modern auth, and WCAG-compliant UX under Agile workflows.
Delivered BI dashboards and governance KPIs; modernized portal to increase staff engagement by 75%.
Produced client-ready BI dashboards with calculated columns, drill-through, and validation frameworks.
Analyzed healthcare regulations and translated legal requirements into actionable technology strategies.
Evidence of communication and teaching — useful signals for collaboration, onboarding, and cross-functional work.
Generative AI Engineer focused on RAG, agents, and fine-tuning. US-based • Remote/Hybrid • Full-time or Contract.
Open Contact Form