Undergraduate at IIT Roorkee building end-to-end LLM applications — RAG pipelines, multi-agent systems, and the full-stack products around them. I care about making AI reliable in production, not just impressive in a demo.
from functools import wraps
def ships(stage: str):
"""Nothing leaves the bench until it's been through here."""
def deco(fn):
@wraps(fn)
def wrapper(*args, **kwargs):
return f"{fn(*args, **kwargs)} → {stage} ✅"
return wrapper
return deco
class LalitYadav:
"""AI engineer — makes LLMs reliable in prod, not just pretty in a demo."""
def __init__(self) -> None:
self.role = "AI / ML Engineer"
self.school = "IIT Roorkee · B.S + M.S (2022–2027)"
self.focus = ["RAG pipelines", "multi-agent systems", "full-stack LLM apps"]
self.stack = {
"agents": ["LangGraph", "LangChain", "Gemini", "Groq"],
"backend": ["FastAPI", "PostgreSQL", "Docker", "AWS"],
"frontend": ["React", "Next.js", "TypeScript", "Tailwind"],
}
self._coffee = float("inf") # ☕
@property
def currently_building(self) -> str:
return "TravelOS — AI agents + CRM for travel agencies"
@property
def philosophy(self) -> str:
return "eval over vibes · ground truth over guessing"
@ships("production")
def build(self, idea: str) -> str:
return f"{idea} → tested"
def __repr__(self) -> str:
return "<Lalit: ship fast · eval faster · build things that actually work>"
me = LalitYadav()
print(me.build("a wild idea")) # a wild idea → tested → production ✅
print(me) # <Lalit: ship fast · eval faster ...>📫 reach me at lalit@ph.iitr.ac.in
🟣 AI/ML Developer Intern — Campayn (Remote · Apr 2026 – May 2026)
- Shipped dashboards, campaign workflows, auth systems and analytics platforms straight from PRDs using Claude Code
- Integrated the Instagram Graph API with fallback scraping for real-time metrics (views, reach, saves, watch-time)
- Automated Instagram/LinkedIn comment-to-DM flows, webhook handling and API integrations — cutting manual ops effort
- Engineered an AI comment generator using a RAG pipeline over 2000+ high-engagement comments, deployed end to end
- Built a job-search platform for career coaches with automated discovery, roadmap generation and pipeline management
- Delivered a CRM, HRMS and salary estimator for lead qualification, plus a company website that improved engagement
Languages & Machine Learning
Generative AI & NLP
Backend & Frontend
Data, Vectors & Infra
| 🚀 Project | What it does | Stack |
|---|---|---|
| Go Issue Agent | Agentic AI that autonomously resolves GitHub issues in Go repos — localizes, fixes, validates against the Go toolchain, and drafts the PR | LangGraph Ollama Go |
| Geospatial Intelligence | Satellite-imagery analysis with U-Net segmentation + YOLOv8 detection and a LangGraph natural-language query agent | PyTorch YOLOv8 FastAPI |
| Fantastic Jobs API | Unified multi-tenant FastAPI backend: job search, resume→title generation, and salary estimation behind one API | FastAPI Groq PostgreSQL |
| Salary Estimator | LangGraph agent that estimates salary from a LinkedIn profile via parallel web search + a ChromaDB knowledge base | LangGraph Gemini ChromaDB |
| LinkedIn Comment Generator | Full-stack Next.js app that writes comments in your voice using a RAG pipeline + fast Groq inference | Next.js Groq RAG |
| Job Dhundo | AI job-search platform for freshers — ATS resume scoring, job matching and a daily roadmap | FastAPI Gemini LangGraph |
- 🏛️ President, Student Affairs Council — led the council, coordinating 6 of 7 General Secretaries across student affairs
- 🎭 Dramatics Section — led a 40+ member team at Inter-IIT Cultural Meets for two years
- 🏔️ Deputy Secretary, Himalayan Explorers' Club — ran finances and logistics for treks across the Himalayas
