DataTalks.Club Build a Strong Career in Data - Lavanya Gupta
15 snips
May 9, 2025 Lavanya Gupta, a Sr. AI/ML Applied Associate at JPMorgan Chase and CMU alumna, shares her journey from software engineering to an AI researcher. She discusses how hackathons ignited her passion for machine learning and her current work benchmarking large language models in finance. Lavanya emphasizes the importance of mentorship and networking for career growth in data science. She explores the evolving roles in the industry, the impact of continuous learning, and the balance between luck and strategy in achieving success.
AI Snips
Chapters
Transcript
Episode notes
Limits of Long-Context LLMs
- LLMs handle easier, shorter long-context tasks well but struggle with long documents in specialized domains.
- Real-world use cases reveal limitations not observed in artificially simplified benchmarks.
Chunk Long Documents for LLMs
- Even with large context window models, chunk long documents for reliable processing.
- Use smaller portions within context size limits to avoid model failure.
Publishing Research in Industry
- Lavanya published a top-tier scientific paper on long context LLM benchmarking while working at JPMorgan Chase.
- Industrial research and publication can coexist as a valuable career experience.