I build agentic and full-stack systems with explicit state, measurable behavior, and interfaces people can understand.
AI / Full-Stack Engineer focused on agentic systems & llm infrastructure, backend APIs, and real-time product systems.
Three projects from the resume, presented through the problem, architecture, interface, and measured outcome.
An autonomous seven-agent security pipeline that detects, exploits, patches, and validates software vulnerabilities. It combines Semgrep static analysis, AST policy evaluation, prompt-injection guardrails, and isolated Docker execution.
A real-time housing and roommate matching product with multi-factor preference scoring, concurrent chat, and low-latency state delivery.
An AI assistant that turns raw soil inputs into crop and fertilizer recommendations for low-connectivity, low-literacy users.
Hands-on software development at high-velocity startups, from autonomous multi-agent systems to backend APIs and GPU compute cohorts.
Orchestration & Guardrails
AEGIS and Raindeer.social show hands-on work with multi-agent roles, workflow state, and validation.
Grounded Recommendations
Agrovers uses FAISS-based RAG and multi-stage validation to improve recommendation reliability by 35%.
Backend to Interface
Kolably and Flately connect backend APIs, real-time state, databases, and responsive product interfaces.
RBI Lending Compliance Engine
RBI lending compliance engine with rapid customer profile retrieval and regulatory rule checks.
LightRAG · Groq API · FastAPI · React · Vite
Explainable Tabular Classification
Explainable tabular classification: XGBoost + LlamaIndex + TabNet audit trails.
XGBoost · RAG · Streamlit · Python · LlamaIndex
Formula 1 Telemetry Analytics
DVA capstone, 7-member team, 744K+ Ergast rows, 10 KPIs analyzing pit window optimization.
Jupyter · Python · Tableau
I'm open to internships and engineering work across agentic systems, backend infrastructure, and full-stack products.