Jashanpreet Singh

Seneca Polytechnic — Toronto

• • Partnered with AMD to build an AI automated test generation system for their firmware C codebase, achieving 96% compile success and 80%+ line coverage on complex, dependency-heavy modules.
• • Built static analysis tooling in Python using Clang to parse C functions and extract dependency graphs, providing the AI agent with codebase context needed for accurate test generation.
• • Developed the AI agent using LangChain and LangGraph to orchestrate test generation across interdependent modules and context retrieval for RAG.

Treasury Board Secretariat, Government of Ontario — Toronto

• • Built an AI-powered Azure DevOps extension in TypeScript that automatically generates test case tickets from requirements, reducing manual test case creation time by an estimated 90%.
• • Proposed and led the project after joining as an automation tester, taking it from initial research through to working prototype.

Next.js, React, JavaScript, Flask

A restaurant social media platform where users can browse restaurants, post comments, and leave ratings. Includes an AI model that suggests restaurants based on past behavior and personalization choices, backed by a Flask REST API for retrieving personalized recommendations.

LangChain, OpenAI API

A voice-powered browser automation AI agent that helps physically challenged users interact with computers hands-free. Built with LangChain and OpenAI's API to translate voice commands into browser actions.

Go, AWS, React

A cloud file-storage microservice built in Go with a React frontend for file uploads and format conversion. Deployed to AWS Fargate with S3 for storage, DynamoDB for metadata, and Cognito for JWT authentication. Includes CI/CD pipelines on GitHub Actions with 80%+ test coverage.

Next.js, Flask

A stock analytics site where users can browse market data and chat with an AI bot to analyze stocks across the American market.