Projects
A few things I've built and contributed to recently.
Dec 2025 — Present
- Fixed a critical logout bug impacting tens of thousands of users on a platform serving 1M+ monthly visits during peak robotics season.
- Improved PWA end-to-end deployment testing, increasing platform reliability.
- Building richer myTBA team-following statistics so teams cut scouting time and avoid multitasking.
March 2025
- Open-source ML model predicting NCAA basketball outcomes from historical team stats.
- Trained with pandas, scikit-learn, and joblib to produce predictions with confidence levels.
- Top-performing bracket ranked #2,092 of ~25M (top 0.01%); 1st in 9 separate groups; 5 brackets in the top 3%.
Promptly — AI Prompt Engineer
AI ATL Hackathon @ Georgia Tech · Honorable Mention
November 2025
- Team-built AI platform for prompt evaluation, benchmarking, and security across multiple LLMs.
- Led the Prompt Injection Detector — trained a custom model that flags jailbreaks and malicious injections with 95%+ accuracy.
- Designed detection logic to reduce data leakage and unsafe prompt behavior; recognized in the Microsoft for Startups track.
Engineer in training,
building real things.
I'm a Computer Science student at Georgia State University with a strong foundation in software engineering, machine learning, and problem-solving — known for combining technical depth with practical thinking to ship efficient, reliable systems.
I'm active in collaborative environments through robotics, mentorship, and technical communities, with a focus on continuous learning and real-world impact. Whether it's a one-line fix that lands for a million users or a model that climbs a national leaderboard, I care about the craft and the result.
Let's build something.
I'm always up for thoughtful conversations — about software, ML, or the next interesting problem.
© 2026 Yaswanth Kandra · Built with Next.js, Tailwind, and Magic UI.