Luka Adzic
SWE @ C4R · Student @ Wharton · Philadelphia
Here's some stuff I've built, things I've worked on, and where to find me online.
Experience
- Software Engineer at Community for Rigor (Philadelphia, PA, May 2026 to present), Scaled an NIH-funded research-rigor platform to 10,000+ students across 7+ institutions with interactive React/Next.js exercises, building the MongoDB backend plus the D3.js peer-comparison tools behind it.
- AI Researcher at The Wharton School, UPenn (Philadelphia, PA, March 2026 to present), Built an explainable pipeline on 40,000+ UK Biobank records comparing a Random Forest baseline to a fine-tuned LLaMA-3-8B, plus a multi-agent pipeline scoring literature evidence for 1,700+ candidate predictors.
- Quantitative Researcher at University of Sydney (Sydney, Australia, May 2025 to August 2025), Built econometric and time-series models in Python and MATLAB to detect speculative price bubbles across 26+ S&P 500 companies, then ran rolling-window backtests across volatility regimes to test signal stability.
Projects
- financial-bubble-detection-dashboard, Real-time financial bubble detection using options data
- maritime-penn-web-app, Penn's first maritime club, founder and president, 100+ members
- algorithmic-trading-system-live-pnl, Live systematic trading system running quantitative trend signals plus NLP sentiment on real capital, up 22% in two months.
- fifa-momentum-tracker, Cracking EAFC/FIFA's dynamic difficulty algorithms through OpenCV and ML-driven pattern recognition.
- gnt-ai, Governs what deployed AI agents are allowed to do at a company. Checks every action against real, PR-approved company rules before it runs, and blocks or escalates anything that doesn't clear.
Socials
lukaadzic ~ %
lukaadzic ~ %