Hi, I'm Tanush
AI Engineer. AI Systems · LLM Engineering · Agentic AI · ML/CV
TM

About

I am an AI Engineer focused on building intelligent systems rather than isolated model demos. My work centers on modern LLM applications, agentic workflows, and production-minded AI engineering. My skills are backed by a strong ML and computer vision foundation and full-stack software engineering background to take ideas from prototype to deployment.

I have hands-on experience across enterprise environments at AT&T and Nokia, independent engineering (Freelance) çfor startups in India, Europe, and the US, and research-driven systems work. I've also filed a patent for Smart Intersections and competed in hackathons spanning AI for Good and Generative AI. Currently I am going deeper into context engineering, agent orchestration, evaluation, reliability, and intelligent automation.

Work Experience

Education

Skills & Expertise

AI engineering is my core, but I build across the full stack - ML, data, web, and cloud.

I'm not a fan of skill bars - they lie. Here's an honest read on where I actually am with the tools I use.

01

AI Engineering

My core focus - building agentic, LLM-powered systems end to end.

LLMs · Agentic AI · RAG Architectures · LangGraph · LangChain · Self-hosted LLMs · Fine-tuning & PEFT · Applied AI · LLM Wrappers · Generative AI · Model Inference · ChromaDB / FAISS

02

Machine Learning & Data

From raw data to trained, deployed models.

Computer Vision · Deep Learning · NLP · Model Development · Object Detection & Tracking · Data Preprocessing · Data Analysis · Model Optimization · Tableau · Python

03

Development, Cloud & DevOps

Shipping full-stack products and getting them live.

Next.js · React · Node.js · FastAPI · Flask · Django · TypeScript · REST APIs · PostgreSQL · MongoDB · Docker · Azure Cloud · Linux · Git / GitHub · Tailwind CSS / ShadcnUI

Languages Python, TypeScript / JavaScript, C, Java, SQL, plus Lua, Julia, R.

CS Fundamentals System Design, OOP, DBMS, OS, Computer Networks, DSA.

Tooling VS Code, Linux, Jupyter, Postman, Figma.

My Projects

Check out my latest work

I've worked on a variety of AI systems, from agentic workflows and RAG products to computer vision pipelines. Here are a few of my favorites.

AgentBreak
GitHubSource

AgentBreak

Red teaming workbench for LLM agents covering prompt injection, tool abuse, privilege escalation, and MCP attack surfaces with automated risk scoring.

Next.js
TypeScript
OpenAI
Prisma
SQLite
Zod
Tailwind
Recharts
Forge
GitHubSource

Forge

Closed loop autonomous coding agent that observes, plans, acts, tests, and opens pull requests with human approval.

LangGraph
OpenAI
FastAPI
Next.js
SQLite
TypeScript
Mnemos
GitHubSource

Mnemos

Persistent memory layer for AI agents with semantic, episodic, and procedural memory plus knowledge graph visualization.

Next.js
TypeScript
SQLite
Knowledge Graph
Agent Memory
AgentLens
GitHubSource

AgentLens

Observability platform for AI agents with span tracing, quality evaluation, and prompt replay comparison.

Next.js
TypeScript
FastAPI
SQLite
OpenAI
MCP Enhanced ERP
GitHubSource

MCP Enhanced ERP

Academic ERP with Model Context Protocol so agents can query institutional data through controlled tools.

Agents
MCP
Claude
LangChain
Python
React
Building Footprint Extraction
GitHubSource

Building Footprint Extraction

Deep learning pipeline that extracts building footprints from satellite and drone imagery into GIS ready outputs.

Detectron2
PyTorch
Computer Vision
OpenCV
Cognital
GitHubSource
Demo

Cognital

Telecom support product with RAG, LLM replies, and human handoff for real operational escalation.

RAG
LLM
Human-in-the-loop
ChromaDB
Flask
React
DeepSeek 7B Fine-tuning
GitHubSource
Notebook

DeepSeek 7B Fine-tuning

Fine tuned DeepSeek 7B with PEFT and LoRA via Unsloth on cloud infrastructure.

DeepSeek
PEFT
LoRA
Unsloth
LLM
GitHubSource

IdeaOS

Agentic product discovery that brainstorms, critiques, researches, and synthesizes ideas into implementation ready blueprints.

Next.js
TypeScript
Supabase
Tailwind
shadcn/ui
GitHubSource

Job Tune

Local resume tailor that rewrites sections against a job description and exports an ATS friendly PDF.

React
Vite
Claude
Express
Tailwind
Viberr
GitHubSource

Viberr

Local memory for vibe coding. Cursor hooks log prompts and edits while a CLI searches and recaps context for the AI.

CLI
Cursor Hooks
TypeScript
Local-first
HireAI
Website
GitHubSource

HireAI

LLM hiring workflow that automates resume screening through interview scheduling.

LLM
Groq
VAPI
React
MongoDB

Hackathons

  • Smart Intersections (Patent)

    2025 · Patent Application No. 202541070871 · AI-driven adaptive traffic signal optimization for real-time traffic flow - filed patent covering the system design and approach.

  • Hack-AI-thon

    2025 · Stella Mary's · Runner Up - AI for Good theme.

  • TechXcelerate

    2025 · BITS Pilani Hyderabad · Finalist among 650+ teams (SaaS track) with HRlytics.

  • 100xEngineers Buildathon

    2025 · Bangalore · Finalist - Generative AI buildathon.

  • Dark Patterns Buster

    2023 · Ministry of Consumer Affairs · Reached the final round.

  • Adeptus - Founding Member / Chief Technical Lead

    2023 - Present · SRM Easwari Engineering College · Technical leadership for the college technical club: ownership of the club website/CMS and systems that support events, newsletters, and community content.

    Website
Contact

Get in Touch

Want to chat about an AI engineering problem? Reach out via Gmail, LinkedIn, or X and I'll respond whenever I can.

GitHub
LinkedIn
X