Experimenting with RL Fine Tuning for LLM mathematical reasoning. Python, PyTorch, HuggingFace

Multi-agent pipeline that reads your resume, discovers real open roles at companies that match your profile, scrapes live recruiter signals, drafts personalized outreach, and autonomously applies

Monitor LLM API usage in real time. Detect cost anomalies, attribute spend to features/teams, and let an LLM explain why a cost spike happened