Summary
Adib Hasan is a research engineer based in New York with 11 years of experience building production-grade ML systems and autonomous agents that operate over long horizons. He combines deep academic work at MIT—where he developed variational pretraining for transformer-based weather models and published on robustness and jailbreak resistance—with hands-on engineering that improved vLLM throughput by 60% and deployed Finance RAG systems using clustering, synthetic data augmentation, and Tree-of-Thought reasoning. His background spans backend engineering, quant research for blockchain portfolios, and production ML optimizations at companies from Facebook to startups, giving him a rare blend of theory and execution. Adib is comfortable optimizing low-level attention kernels and designing high-level agent orchestration (MCP for Jupyter automation), so he moves models from prototype to scalable tooling. Colleagues describe him as pragmatic yet research-driven, and he often applies mitigation techniques (like moderate pruning) as practical regularizers rather than purely theoretical fixes.
11 years of coding experience
5 years of employment as a software developer
Master of Engineering - MEng Computer Science, Master of Engineering - MEng Computer Science at Massachusetts Institute of Technology
English, Bengali, Hindi, Spanish