Nikhil Mourya - ML Engineer · Multi-Agent Systems

About

I'm Nikhil Mourya. I build ML systems where the math decides when to stop, not the model's confidence. Retrieval scores over self-reported certainty. Pruned weights over bloated checkpoints. Shipped code over benchmark screenshots.

BIT Mesra, studying what doesn't fit in a lecture hall. Right now: questioning whether your retrieval pipeline actually needs that vector DB, or just needs a better question.

I'm always maintaining unbroken eye contact with that single code block like we have beef.

Model Compression

Pruning and LoRA are quiet admissions that full fine-tuning is often billing for capacity you never needed. I like models that shrink without forgetting what they're for.

efficiency

NLP & Text Systems

Summarization, classification, multi-agent pipelines where fluent isn't the same as faithful. Production NLP is mostly telling confident hallucinations they can't sit with us.

production

Multi-Agent Systems & MLOps

LangGraph orchestration, eval loops that catch what the LLM won't admit, FastAPI + Docker + AWS, the glue between a Jupyter notebook and someone else's pager.

infra

Competitive Programming

Codeforces Specialist, 1400+. Graphs have a way of humbling you on a schedule; the upside is you stop trusting clever one-liners without proof.

cf · 1400+

ongoing

Specialist Codeforces 1400+

Still grinding rated rounds, the graphs are optional; the ego damage isn't. Proof you can think with a clock breathing down your neck.

peak: 1400+ · still climbing

2024

Finalist SIH - Hospital Mgmt System

99.5% uptime with real models in the loop, the 0.5% was character development. Backend stayed polite even when the night shift wasn't.

99.5% uptime · production ML

2023

Finalist IIIT Delhi - ResNet50

94% accuracy after grid search stopped me from brute-forcing the hyperparameter void. Sometimes the boring search is the clever move.

94% acc · −30% train time

Skills & Arsenal

OpenAI
API

AWS

Lang
Graph

LoRA
PEFT

RAG

FAISS

ChromaDB

RAGAS

Lang
Smith

Groq API

REST

tree
sitter

asyncio

Fine
Tune

Prompt
Eng

Model
Comp.

Serving

+

Projects

nikhil@portfolio:~/projects $ ls -la

~/projects/flagship_projects

vLLM Inference Engine

98.5% TTFT reduction

GitHub

"Can we eliminate prefill-blocking-decode stalls for optimal continuous batching?"

An 11-phase LLM serving engine implementing continuous batching, paged KV cache (block allocator + physical tensor pool), and chunked prefill. Cut mean TTFT under 4-way concurrent load by 98.5% (1,418ms → 20.9ms) and wall-clock latency by 26.5% vs. naive sequential serving.

Diagnosed: KV-cache pool-reconstruction bug adding 1.1ms/token overhead; fixed by keeping HuggingFace past_key_values live on-sequence during decode.

Continuous Batching Paged KV Cache Chunked Prefill

Python PyTorch HuggingFace LLM Serving

HiveMind

GitHub

"Can a research system know its own answer isn't good enough, and fix it?"

Production-grade autonomous research system with a 5-agent sequential pipeline (Planner → Researcher → Critic → Writer → Evaluator). Orchestrator loops autonomously until a deterministic confidence threshold — calculated as mean retrieval score, not LLM self-report — is met.

Infra: asyncio.Lock single-threaded LLM execution · Docker Compose (FastAPI + Redis + ChromaDB + OpenTelemetry) · SSE streaming

5-agent pipeline autonomous eval loop deterministic confidence scoring

Python FastAPI ChromaDB Redis OpenTelemetry Groq Ollama Docker

CodeLens

GitHub Demo

"Can you search a codebase by intent, not keywords, fully offline?"

Fully offline VS Code extension that indexes codebases using AST-based chunking via tree-sitter and 768-dim embeddings via Ollama. Natural language query → ranked semantic results → one-click jump to exact line. No internet, no API keys.

9K+ chunks indexed sub-10ms ANN query 100% offline

Python TypeScript FastAPI tree-sitter Ollama VS Code API Docker

~/projects/other_work

01

AttentionIsALLICode

Not "used a framework." Actually from scratch.

Full architecture Multi-head attention Custom training loop

PyTorch Transformers NLP From Scratch

02

Vectorless RAGs

Vector databases are the default. I wanted to know if the default was actually necessary.

Zero embeddings Zero vector DB Full retrieval

Python Ollama LLaMA 3 RAG Streamlit

03

Second Brain Debugger

A senior engineer code-reviewed my brain. Six stages. Real AI. No affirmations.

6-stage pipeline SSE streaming Multimodal input

Next.js 14 TypeScript Mistral-7B Whisper Stable Diffusion

04

DermaVision

7 skin lesion classes. 57:1 class imbalance. Focal Loss said no problem.

HAM10000 dataset Grad-CAM XAI ONNX export

EfficientNet-B3 PyTorch FastAPI Next.js 14 Docker

05

PEGASUS + LoRA · Efficient Summarization

99.8% param reduction

Fine-tuned a 767M parameter model using only 1.57M trainable parameters via LoRA. Full fine-tuning produced incoherent outputs on unseen domains. LoRA didn't.

27× faster training 767M → 1.57M

PyTorch HuggingFace PEFT LoRA NLP

06

Pruned U-Net · Biomedical Segmentation

97.3% param reduction

How much of a segmentation model is actually load-bearing? IIT Kharagpur research collaboration.

92% FLOPs reduction IoU > 0.95 on MoNuSeg

PyTorch U-Net Model Pruning Computer Vision

// ventures

Founder

DevPath

Structured learning for developers who are tired of tutorial hell.

Growingcurated roadmaps

Dailyfocused tasks

Freeto get started

Visit DevPath ↗

Experience

2025

Deployed Software Engineer - ML @ HireBuddy

Built and deployed a resume-JD matching system using RAG and Transformers, completely replacing legacy regex parsers.

Designed an NLP ingestion pipeline processing 10k+ resumes/day for automated shortlisting.
Increased shortlist accuracy by +35% and reduced chain latency by 28%.
Shipped robust REST APIs using FastAPI and Docker on AWS.
Set up a deterministic evaluation stack using RAGAS and LangSmith to track and improve metrics.
Improved candidate-role matching, boosting overall user engagement by +40%.

RAG · Transformers · FastAPI · AWS · LangSmith

If you've made it this far,
you might as well say Hi.

Email tsmftxnikhil14@gmail.com

LinkedIn nikhil-mourya

GitHub TryingtobeingNikhil

Twitter / X @GonnabeNikhil

Instagram @not_nikhil14