Hi, I'm Tejdeep
I build backend systems for AI products: APIs, data flow, workflow state, and human approval.
TP

About

I’m a Computer Science graduate from George Mason University with experience building backend services, cloud data infrastructure, and AI workflows. At Wi-Tronix, I built Python and SQL ingestion pipelines for 50,000+ daily rail camera images and AWS workflows managing 5TB+ of ML data with S3, Apache Iceberg, Athena, and EKS. At Miraivant Advisory, I built 10+ FastAPI and MongoDB endpoints, a multi stage candidate matching service, and CI/CD pipelines for AWS EC2.

My recent projects include a FINRA compliance RAG system with hybrid pgvector and full text retrieval, citation validation, and abstention, plus LangGraph workflows with persisted state, retries, and human approval. I’m targeting full time software engineering roles across backend, data platforms, cloud systems, and applied AI.

Work Experience

Education

Skills

Python
TypeScript
React
Next.js
FastAPI
PostgreSQL
MongoDB
Redis
AWS
Docker
Kubernetes
LangGraph
RAG
Java
My Projects

Check out my latest work

I've worked on a variety of projects, from simple websites to complex web applications. Here are a few of my favorites.

GitHubSource

Replic

A multi-tenant AI marketing agent that monitors social platforms, scores opportunities, generates on-brand content, and routes it through human approval before publishing. Built with safety controls and an auditable activity trail.

Next.js
TypeScript
FastAPI
Supabase
PostgreSQL
Redis
Composio
GitHubSource

MoveOS

An asynchronous relocation agent that discovers, normalizes, and deduplicates housing listings, evaluates budget, commute, and lease constraints, and persists ranked decisions and workflow state in MongoDB.

React
TypeScript
Python
FastAPI
LangGraph
MongoDB
GitHubSource

FINRA Compliance RAG

A source-grounded question-answering service for FINRA rules. It uses section-aware ingestion, hybrid pgvector and full-text retrieval, evidence validation, precise citations, and explicit abstention when the corpus cannot support an answer.

Python
FastAPI
React
PostgreSQL
pgvector
Docker
Gemini
GitHubSource

TennisIQ

A computer-vision platform for college tennis coaches that turns match footage into court detection, ball and player tracking, point segmentation, visual heatmaps, and actionable coaching insights on GPU infrastructure.

Python
FastAPI
Next.js
YOLO
OpenCV
Modal
Beyond Projects

Research, teaching & recognition

I also study model internals, teach programming, and build under competition constraints—experiences that sharpen how I reason, communicate, and ship.

Research

Research Assistant · LLM Watermarking

George Mason University

Prototyping token-level sampling changes in vLLM and benchmarking GPU activation capture with a focus on preserving inference throughput.

Teaching

Teaching Assistant · Java & Python

George Mason University

Mentored 50+ students through labs, code reviews, and one-on-one guidance in object-oriented programming, data structures, and systematic debugging.

Recognition

7-Time Dean’s List Honoree

George Mason University

Recognized for academic performance across seven semesters of the Computer Science program.

Hackathons

2-Time Hackathon Finalist

HackPrinceton · HoyaHacks

Reached the finals at HackPrinceton and HoyaHacks after building working products under tight team and time constraints.

Contact

Get in Touch

Have a role, project, or idea worth discussing? Send me an email, and I'll get back to you.

GitHub
LinkedIn