I'm a Senior Data Scientist at Optym, where I work on optimization and machine learning. Before that I was a Backend Engineer at Gojek and a Platform Engineer at Mailmodo. I studied at IIT (BHU) Varanasi.

may the code you write bring you happiness

• Combinatorial optimization. Exact methods, bounds and certificates, column generation and pricing problems. I like results that a machine can re-verify from scratch.
• Applied machine learning. Models that sit next to an optimizer and make it better, not models for their own sake.
• Backend systems. Years of shipping Python and Go services at scale, which keeps my research code honest.

A computational attack on the asymptotic competitive ratio of online bin packing, with the accompanying paper.

• Improved the verified upper bound of the Advanced Harmonic algorithm from 1.5782896 to 1.5782147, through a re-optimised proportion table and per-scenario optimal dual weights.
• Proved two theorems showing the 2018 lower-bound construction cannot be improved by adding new item batches, and closed the remaining window with machine certificates.
• Every number is exact rational arithmetic, and every claimed bound ships with a certificate that an independent checker re-verifies. Start at docs/verification-guide.md if you want to check rather than read.