# Sahaj Chhabra

**Software Engineer** — Canada

> Full-stack engineer building AI-powered video and media systems at scale, from thumbnail generation pipelines to verticalization APIs and distributed encoding infrastructure.

Career Archetype: **Launchpad Trailblazer** (HS-MG) — One of the first Launchpad Trailblazers on Saywise

## Links

- LinkedIn: https://linkedin.com/in/sahaj0312
- GitHub: https://github.com/Sahaj0312

## Experience

### Software Engineer, Quickplay (2025-01 – 2026-09)
Toronto, Ontario, Canada · Built an end-to-end AI thumbnail generation pipeline combining computer vision for frame extraction with LLM-powered captioning for thousands of video assets. Designed and shipped a Verticalization API using YOLO object tracking to convert 16:9 videos to 9:16 vertical shorts at scale for the Pilipinas Live app. Engineered a Golang microservice to automate n8n workflow deployments, eliminating 90% of manual migration work. Architected a distributed HLS encoding system in Go generating adaptive bitrate streams for 20,000+ video assets using multithreading and Pub/Sub messaging.

### Co-Founder & CTO, Instagrad (2024-09 – 2025-12)
Vancouver, BC · Co-founded an AI startup to streamline photo studio operations and secured a partnership with Vancouver's largest studio, gaining access to 10+ petabytes of image data to train a proprietary on-premise model for AI-generated graduation composites. Built an automated face-matching pipeline that ingests thousands of graduation photos and clusters them into per-student folders using facial recognition, replacing a manual sorting process.

### Software Engineer Intern, Quickplay (2024-06 – 2024-08)
Toronto, Ontario, Canada · Implemented a GraphQL-to-SQL translation engine with nested query support and batch resolution to eliminate N+1 query overhead. Led testing efforts across multiple microservices with unit and integration tests in Go using table-driven patterns, increasing SonarQube test coverage from 14% to 97%. Collaborated with three cross-functional teams to build an internal tool converting legacy XML metadata files to JSON compatible with any web-based CMS.

### Software Developer Intern, BMO (2022-05 – 2022-08)
Toronto, Ontario, Canada

## Education

### Bachelor's degree in BUCS, The University of British Columbia (2019-09 – 2024-06)

### International Baccalaureate, GEMS Education (2005-01 – 2019-01)

## Projects

### Pixitt
Camera roll cleaning app with Tinder-like interface · Built a camera roll cleaning app using SwiftUI with a Tinder-like interface for deleting duplicate or low-quality photos. Implemented duplicate detection using perceptual hashing and on-device ML with MVVM architecture, asynchronous media loading, and optimized memory management to handle large photo libraries efficiently.

### InvestorAI
Personal finance tool using LLMs for portfolio analysis · Built a personal finance tool that uses LLMs to analyze stock portfolios and generate insights based on financial documents and market data. Created embedding pipelines using pgvector for context-aware retrieval with RAG architecture, deployed to Azure Kubernetes Service.

### Fit@Home
Real-time AI personal trainer with video streaming · Built an AI personal trainer for nwHacks that streams video over WebRTC to a Python backend using MediaPipe to track 33 body landmarks and calculate joint angles for real-time exercise form validation.

## Skills

Python (Programming Language), JavaScript, MySQL, Java, HTML, Cascading Style Sheets (CSS), Go, Swift, FastAPI, Flask, React, SwiftUI, LangChain, OpenCV, MediaPipe, Docker, Kubernetes, GCP, AWS, Azure, Git, PostgreSQL, n8n, LLMs, RAG, YOLO, Computer Vision, pgvector

## Credentials

- **AI Fundamentals Certificate** — UnlockAI: Learn AI Daily (issued 2026-09)

---
Source: https://saywise.com/member6268 (last modified 2026-10-05T22:49:49.013Z)