About Mohit Goyal

AI Engineer · LLM Agents, RAG, Evals

Who I am

I'm Mohit Goyal. I am an AI engineer in New Delhi, India. I build LLM agents, RAG systems and agent evals, and I co-authored GRIT, a parameter-efficient fine-tuning method that trains 46% fewer parameters than LoRA. Most of my work lives at the edge of research and shipping: reliable systems, not just impressive demo windows.

Track record so far: 3 publications / 15x national hackathon winner / research-to-prod AI systems.

What I focus on

Most of my time goes into Agentic AI, RAG infrastructure, Applied LLM research, and MCP systems. Day to day, that means working across Python, PyTorch, FastAPI, MCP, LangGraph, Next.js, moving between the model layer and the product layer on the same project.

Currently

Right now I'm at Lenskart, and here's what I actually work on there.

AI Engineer at Lenskart

Aerocity, New Delhi, India (On-site) · July 2026 – Present

  • Building three internal AI systems on Lenskart's AI Engineering team: a hiring intelligence platform, a CEO intelligence brief, and a competitor creative-intelligence platform.
  • Made candidate searches survive crashed workers and provider outages. A lease-and-heartbeat state machine over Postgres reclaims a dead worker's in-flight work, and per-provider failure classification falls back across 4 LLM providers mid-search.
  • Made candidate ranking explainable and repeatable. A deterministic 4-key comparator (score, evidence coverage, confidence tier, name) replaced one opaque LLM score, and identity dedup sends ambiguous matches to human review instead of auto-merging them.
  • Made sure no claim reaches the CEO brief on one weak source. An evidence layer admits a claim only on 1 primary or 3+ independent origins above 0.75 match confidence. Built end to end: 40 source connectors, a 9-tool MCP server backing 18 curation skills, and a measured cost of $1.09 per curation session.
  • Replaced a creative-analysis pipeline that fabricated 27% of its rows (78 of 287). In the new 7-layer FastAPI and Next.js platform, a deterministic guardrail checks every generated figure against tool evidence instead of trusting an LLM judge.
  • Raised LLM-judge agreement with human labels from 54% to 89% (27 held-out labels in the latest batch) with a tiered judge (deterministic checks, then a classifier, then LLM review), planted-error tests, an overfit audit and a regression gate.

Earlier

Before that, I worked as a Software Engineer (Part-time) at C3alabs, and as a Full Stack Developer Intern (Part-time) at ArmorIQ, and as a ML Researcher at RAAPID INC, and as a AI Agent Developer Intern at Kartavya Technology. The full breakdown of what I did at each is on the work experience page.

Research & publications

I also do applied ML research, mostly on efficient fine-tuning and explainable, trustworthy systems. So far that has produced 3 publications:

  • GRIT: Geometry-Aware PEFT with K-FAC Preconditioning, Fisher-Guided Reprojection, and Dynamic Rank Adaptation — arXiv preprint (2026). Read the paper
  • AoT-BTS: An Explainable Age-of-Trust Based Behavioral Scoring Framework for Real-Time Zero Trust Enforcement using SDN — Research Square (2026). Research Square
  • A Comprehensive Review of Microsoft Semantic Kernel: Architecture, Capabilities, and Research Directions — IEEE DELCON (2025). IEEE / DOI

What drives me

Ikigai is a Japanese idea for the reason you get up in the morning: the overlap of what you love, what you're good at, what the world needs, and what you can be paid for. Here's mine, as a person, not just an engineer.

Travel, New Ideas, Music & AnimePeople With Common SenseAgentic AI, RAG, ML EngineeringBuilding Production AI Systems
Portrait of Mohit Goyal
What I love
Traveling, exploring new ideas, collaborating with people, music, and media like anime, manga, and web series.
What I'm good at
Agentic AI, RAG infrastructure, applied LLM research, and MCP systems, with 3 publications and a 15x national hackathon win record to show for it.
What the world needs
More people who lead with common sense.
What I can be paid for
Building agentic AI systems and ML infrastructure in production, at Lenskart, C3alabs, ArmorIQ, RAAPID, and Kartavya.

More about me

  • Journey — a timeline of what I learned and shipped
  • Certificates — programs and achievements
  • Gears — the devices and tools I work with
  • Setup — my VS Code configuration and extensions
  • Blog — write-ups on agents and ML research
  • Projects — case studies of what I have built

Quick answers

Who is Mohit Goyal?
Mohit Goyal is an AI engineer in New Delhi, India. He builds LLM agents, RAG systems and agent evals at Lenskart, and is a final-year B.Tech Computer Science student at Bharati Vidyapeeth's College of Engineering with a CGPA of 9.24, graduating in June 2027.
What does Mohit Goyal work on at Lenskart?
At Lenskart, Mohit Goyal builds three internal AI systems: a hiring intelligence platform, a CEO intelligence brief, and a competitor creative-intelligence platform. On the creative-intelligence agent he raised LLM-judge agreement with human labels from 54% to 89% (27 held-out labels).
What has Mohit Goyal built?
Mohit Goyal built Soundstage, which turns a voice brief in Hindi, Hinglish or English into an on-brand 9:16 reel (one run scored 91.8/100 in 3 minutes 36 seconds); Outfox, competitor research where every claim links to its public source; AgentForge, a coding-agent harness on PyPI; and MemexLLM, a RAG platform with sub-500 ms search.
What research has Mohit Goyal published?
Mohit Goyal has 3 publications: AoT-BTS (first author, Research Square preprint, 98.31% accuracy for zero-trust enforcement), GRIT (co-author, arXiv:2601.00231, a LoRA variant that trains 46% fewer parameters than LoRA), and a peer-reviewed review of Microsoft Semantic Kernel at IEEE DELCON 2025.
Is Mohit Goyal open to new roles?
Yes. Mohit Goyal is open to remote forward deployed engineer (FDE), contract or full-time AI engineering roles. He is based in India (IST). The contact page is the fastest way to reach him.

Get in touch

If you want to work together, talk about a project, or ask about the research, the fastest way to reach me is the contact page. You can also find my resume at /resume.

“A man who is master of patience is master of everything else.”

~ George Savile

© 2026 Mohit Goyal