01
Sentinel
Guardrails and quality gates for agentic AI systems — validating LLM output before it ships.
▮ PORTFOLIO — 2026 · AGENTIC SYSTEMS
I build agentic AI systems — guardrails, RAG pipelines, and the orchestration that holds them together.
▮ SEC.02 — THE STORY
CH.01
First programs, first bugs, first all-nighters. I learned by building, breaking and rebuilding.
CH.02
LLMs that don't just answer — they act. I build the guardrails, retrieval and orchestration that make them reliable.
CH.03
Everything below runs from a real repository. The commit graph tells the rest of the story.
SEC.02 — ABOUT
I'm Aditya — a computer science student at MIT (class of ’27) who builds and ships. Most of my time goes into agentic AI: systems where LLMs don't just answer, they act — retrieving, checking and routing work through pipelines I design.
In between I've built full-stack products, OCR pipelines for messy real-world documents, and enough side projects to keep my commit graph honest. Everything on this page links to code you can actually read.
SEC.03 — WORK
Selected projects — every link points at real code on GitHub.
01
Guardrails and quality gates for agentic AI systems — validating LLM output before it ships.
02
Retrieval-augmented generation with an agentic loop — fetch, re-rank, answer, then check the answer.
03
AI tailoring for job applications — resumes and cover letters tuned to each role.
04
Document pipeline for healthcare records — extraction, cleanup, structured output.
SEC.04 — STACK
SEC.05 — EDUCATION
B.Tech, Computer Science & Engineering
2023 — 2027
Coursework across algorithms, systems and machine learning — applied directly in the projects above.
SEC.06 — CONTACT
Building something with agents in the loop? My inbox is open.