Hey, I'm Abhinav. Most of my work revolves around writing Python and systems code for autonomous multi-agent workflows, experimenting with edge inference runtimes, and curating clean datasets for post-training alignment.
I believe in local-first development: writing software that runs, tests, and validates reliably on personal hardware before relying on cloud clusters.
“Nothing is impossible when written in code.”
A usable, modular framework in Python for multi-agent software development. Handles planning, code generation, compilation checks, and sandboxed validation loops.
A lightweight runtime SDK designed to link local and edge machine LLMs into a coordinated compute mesh, reducing reliance on centralized API endpoints.
A curated open-source Supervised Fine-Tuning (SFT) corpus formatted in ChatML. Filtered to remove benchmark overlap and personal identifiable information for reliable behavioral alignment.
If any of my open-source libraries or datasets have helped your projects, crypto contributions directly fund local testing hardware, electricity, and compute costs for training runs.
0x2A1eEC07BC7DdFD09A54F160ba13748BB9f3c2Ae
Whether you want to discuss multi-agent architectures, collaborate on distributed inference runtimes, or talk about open datasets, feel free to reach out directly.