Building practical AI systems with a focus on GenAI, LLM security, automation, and reliable ML engineering.
- AI threat modeling & LLM security — mapping vulnerabilities to MITRE ATLAS and the OWASP LLM Top 10, and building guardrails around production GenAI systems
- RAG & LLM benchmarking at Siemens Healthineers — designing evaluation frameworks and inference-optimization pipelines for LLMs in a regulated healthcare setting
- Multi-step agentic workflows with LangGraph, orchestrated through Azure CI/CD
- A Snowflake-integrated QA bot for automated testing and data validation
- Sharpening low-level design fundamentals through a structured, self-directed study program
Siemens Healthineers — GenAI / Automation Engineering
Worked on enterprise AI initiatives involving:
RAG evaluation and benchmarking LLM inference and evaluation workflows AI-assisted quality and process automation Multi-step agentic workflows Data extraction and validation pipelines AI security and threat modeling Cloud-based orchestration and automation
Note: Some of this work was developed within Siemens Healthineers' internal environment and therefore isn't publicly available on GitHub.
| Project | What it does |
|---|---|
| llm-security-gateway | Enterprise-grade LLM security gateway (FastAPI) with pluggable scanners, policy-as-code, RBAC, and OWASP LLM Top 10 mapping — Dockerized with a GitHub Actions CI pipeline |
| architect-rl-env | A reinforcement-learning environment built on a two-phase consultant→architect decision engine (constraint elicitation, trade-off reasoning, contradiction detection, ADR-style export) — features a three-layer state structure, multi-path scoring, and adversarial "chaos modes" |
| ForensAI | Open-source AI pipeline for crime-scene analysis: NLP for FIR parsing, RAG for case-similarity search, and case profiling |
| ride_sharing_demand_autogluon | Ride-sharing demand forecasting tackled via AutoGluon, from a Kaggle challenge |
- 🎓 Delegate, Harvard Project for Asian and International Relations (HPAIR)
- 📜 Patent credit from a research assistantship at VTU
- 🏎️ Top 500 globally, AWS AI/ML Scholarship, via PPO-based reinforcement learning in the AWS DeepRacer Student League
- 🌱 Open-source contributor — GirlScript Summer of Code & Hacktoberfest
- ✍️ Writes on Medium, documenting the engineering work behind these projects
Languages & Core
AI / ML
Infra & Data
Open to opportunities in GenAI engineering, LLM security, and applied ML — let's connect.