Senior AI Alignment Specialist - TechNova AI Solutions
Lead alignment efforts for a large language model (LLM) across coding, math, and physics domains, ensuring outputs meet safety and logical consistency requirements. Developed domain-specific annotation guidance and reduced hallucination rates for financial analysis tasks using rigorous evaluation. Collaborated with engineering teams to debug generated code and validate adherence to best practices and security standards. • Evaluated 500+ prompts daily for logical consistency and safety compliance. • Reduced hallucination rates in stock market projections by 18% through guideline and evaluation improvements. • Partnered with engineers to debug Python and C++ code and ensure security and best-practice alignment. • Mentored a team of 10 junior annotators on ATC terminology and safety protocols to improve accuracy scores to 99.5%.