AI Engineer & Generative AI Researcher | LangGraph · RAG · LLM Evaluation · Multi-Agent Systems · Python · SQL · BI

I don't just use AI - I build systems that make it reliable, evaluated, and production-ready. 🚀

LangGraph · RAG · GraphRAG · LLM Evaluation · Multi-Agent Systems · Text-to-SQL · Python · SQL · Pinecone · Neo4j · Snowflake · dbt · Power BI · Tableau

📄 Live Demo - ClariRAG  |  📄 Live Demo — MetricMind

• Agentic RAG pipelines with hybrid retrieval, citation validation, and sufficiency guardrails - systems that say "I don't know" instead of hallucinating
• Governed text-to-SQL agents scoped to certified dbt metrics, so the LLM cannot invent numbers that don't exist in the semantic layer
• GraphRAG systems combining Neo4j knowledge graphs with vector retrieval for multi-hop relationship reasoning flat vector search can't do
• Multi-agent evaluation systems that audit AI output before it reaches a human, with schema-validated contracts between agents
• LLM observability stacks scoring faithfulness, cost, latency, and hallucination rates with LangSmith, Langfuse, Ragas, and Grafana

AI/ML: LangGraph · LangChain · Anthropic Claude · RAG · GraphRAG · Pinecone · BM25 · Ragas · DeepEval · LangSmith · Langfuse · Cross-Encoder Reranking
Graph & Data: Neo4j · Cypher · pgvector · DuckDB · Snowflake · dbt · Azure AI Search · ETL · Pandas · NumPy
BI: Power BI · Tableau · Looker Studio · DAX
Infra: FastAPI · FastMCP · Streamlit · Redis · Prometheus · Grafana · Docker · GitHub Actions · Vercel · Railway · Fly.io

The LLM cannot invent a metric that doesn't exist — every answer traces back to a certified dbt model

Business teams wait 3–7 days for analysts to answer questions like "what is 30-day retention for the EU cohort, adjusted for refunds?" The deeper problem is metric drift: "active user" means something different across five dashboards. MetricMind solves both.

• 5-node LangGraph pipeline: Intent Classifier → 3-layer Guardrail → SQL Generator → DuckDB Executor → Response Node, with automatic self-correction on broken SQL
• 3-layer guardrail: PII regex + SQL injection regex + metric allowlist via Claude Haiku — bad queries rejected for $0.0003 vs $0.006 for full pipeline (20x cheaper)
• Governed semantic layer: 6 certified metrics in a JSON catalog; the agent is physically scoped to only those — no hallucinated numbers, no schema drift
• dbt Core: 4 staging models + 4 mart models (DAU, cohort retention, revenue, funnel) with 40 dbt tests catching 150+ dirty rows before any reach a prompt
• 100% eval accuracy on 50-question golden set scored via sqlglot AST comparison · prompt caching on the 3,000-token metric catalog (90% cache hit rate, ~$0.006/query avg)
• Dual anomaly detection: 3-sigma rolling window + Prophet, with HITL commentary approval before publishing
• Full LLMOps stack: LangSmith traces every node · Prometheus scrapes FastAPI every 15s · Grafana dashboard (latency, cost, guardrail rejections) · Tableau Public dashboards
• React + Vite frontend · FastAPI backend on Railway · Vercel deploy · Docker Compose local stack

🔗 Repo · Live Demo

Clinical knowledge retrieval that shows its work, and knows when to stay quiet

Every claim is tied to a page number. Every citation is validated before it reaches the user. If the answer isn't in the corpus, the system says so — instead of guessing.

• 5-node LangGraph pipeline: Analyser → Expander → Hybrid Retriever → Sufficiency Judge → Generator, with a conditional retry edge when context falls short
• Hybrid retrieval: BM25 (exact clinical terminology) + Pinecone dense vectors, fused with RRF and reranked by a cross-encoder on the top 20 candidates
• Retrieval hit rate improved from 58% → 81%; hallucinated citations reduced to zero via hard guardrail validation
• Ragas faithfulness 0.86 · LangSmith node-level tracing · FastMCP server (usable from Claude Desktop) · React + Vite frontend on Vercel
• Corpus: 5 WHO clinical guideline PDFs · 299 pages · 1,911 chunks

🔗 Repo · Live Demo

Reads an invoice image → extracts clean structured data → answers multi-hop supplier questions that vector search can't

Finance teams manually key 50,000+ invoices/month at ~$3.50/invoice. Pure vector RAG can't answer "which suppliers tied to delayed Q3 POs also had quality complaints in the past 18 months?" — because it has no concept of graph structure. LedgerLens solves both halves.

• Claude vision extraction with per-field confidence scoring (0.0–1.0) and automatic human-review routing for low-confidence documents — no OCR pre-processing required
• Neo4j knowledge graph maps Supplier → Invoice → LineItem → PO for relationship reasoning across entities
• LangGraph GraphRAG agent: plan → retrieve → traverse → answer, returning the full Cypher traversal path as an auditable explanation
• 84.2% field extraction accuracy · 99.7% cost reduction vs $3.50/invoice manual baseline (~$0.008/invoice with Claude Sonnet)
• DeepEval/RAGAS eval harness · Langfuse span-level tracing + per-document token cost · FastAPI + React UI · Docker + Fly.io deploy

🔗 Repo

Role-aware technical documentation intelligence with LLM-as-judge evaluation and a CI/CD eval gate

Engineers at semiconductor manufacturers spend 2–3 hours per shift searching thousands of pages of machine manuals, fab process specs, and compliance guidelines. A wrong answer can stop a production line.

• 5-agent LangGraph pipeline: Query Understanding → Privilege Check → Hybrid Retrieval → Citation Grounding → LLM-as-Judge Evaluation
• RBAC enforced at the retrieval layer — server-side access filtering, not just at the API boundary
• Dual LLM architecture: Azure OpenAI GPT-4o for generation, Anthropic Claude as a separate judge model — keeping generation and scoring fully independent
• HITL gate: confidence < 0.60 routes to human review instead of shipping a weak answer
• 65 passing tests across chunker, loader, search, agent, and pipeline layers · 30-question tiered golden eval dataset (factual, procedural, multi-hop) · CI-gated eval regression on every push
• Prompt versioning via JSON config registry · 3 chunking strategies with ADR documentation · Full operational runbook

🔗 Repo

Multi-agent pipeline that turns a plain-English hiring brief into an evaluated, export-ready candidate slate

Most AI pipelines stop at "the model returned valid JSON." SearchIQ treats that as the easy 10% of the problem.

• 4-agent pipeline: Market Mapper → Profile Generator → Critic Agent → Exporter
• Critic agent scores every profile against 5 structured criteria (title match, accountability ownership, credential specificity, brief-specific fit, domain translation risk) before any slate ships
• Schema-validated JSON contracts between agents; failed validation triggers a corrective retry with the error fed back into the prompt
• Multi-provider: Claude Sonnet / Haiku, GPT-4o, Gemini — swappable via a single config file
• Versioned prompts with v1 limitations documented inline — the iteration reasoning is visible, not just the final result
• Streamlit UI · Google Sheets export with CSV fallback

🔗 Repo

• Building production-grade agentic AI evaluation frameworks
• Deepening expertise in LLM observability (LangSmith, Langfuse, Ragas, DeepEval, Grafana)
• Expanding into GraphRAG, multimodal AI, and governed analytics for enterprise use cases
• Targeting AI Engineer / AI Analyst roles at the intersection of AI systems and analytics

• 🎓 MS Information Systems — San Diego State University (GPA: 3.7)
• 🎓 B.Tech Computer Science & Business Systems — DY Patil College of Engineering (GPA: 3.8)

Most people ask "does the AI return an answer?"
I ask "is the answer faithful, grounded, and verifiable — and what happens when it isn't?"

⭐ If my work is useful, feel free to explore the repos!