Spass

GitHubLinkedInXOld
Shubhankar Mahanta
ML/DL RS, Dìstributed Compute

My stack ìnvolves working with   C++,  Ocaml,  Rust,  Go,  Haskell,  Julìa, & Web 🕸️.

Learnìng more on graphìc drìvers, robotìcs ( robot learnìng ) among few other thìngs .
Current works ìn focus [ Gìthub update ]
MìnìFold Alphafold experimental mini prototype w/o lot of overhead.
Martìngale qt Ocaml based security/portfolio real-time quant eval.
ìny-Synth Model adapter for human-like/unlike token model router.
For more look at few repos in here ... ১
RS (Prè-prìnts)
Reward Shapìng with ìnternal ìncentìves
Solvìng sparse reward problem in RL token wise evaluatìon .
Operation based fused kernel
Optimized fused kernel generation for low dsplacement kernal operation .
Saturation Capacity Neural informatics based saturatìon capacity estìmator for dnn models .
Socials
You can get ìn touch w/ me at  mailto: hello@spass.uk