I work with models every day.
From my first C++ class in 11th grade to building AI systems at scale, I've been driven by solving real-world problems through technology.
What I've done so far. I'm pursuing my Master's in Data Analytics Engineering at Northeastern University (GPA: 3.92/4.0), building on my B.Tech from Manipal University Jaipur. As a Software Development Engineer at Agentdesks (Radius Agent subsidiary), I developed AI chatbots and RAG systems for 70,000+ real estate professionals, while architecting CI/CD pipelines that cut build times by 57%. For my capstone I shipped the Workhuman HR Analytics Capstone, an end-to-end HR analytics pipeline that pulls qualitative insight out of employee recognition data; routing bulk NER and extraction to a fine-tuned local Llama-3-8B instead of a pure LLM API cut model serving cost 10x with no loss in output quality. I also contribute to open source outside my own projects, mostly in the AI and research tooling space.
What I'm building right now.
Much of my time goes to private repos. In public it goes to **[Coo](https://github.com/abhinavaditya811/coo)**, a kernel that turns a natural-language request into an ordered plan of pre-defined capabilities and runs them on a Mac over HTTP. The model only picks capabilities and fills in their parameters and never writes the code that runs, and an always-on edge still gives an honest answer while the Mac is asleep. I also keep **[Settl](https://settlnow.vercel.app)** going: it chases overdue invoices for freelancers, and its decision core is deterministic and runs offline.Beyond tech, I bring leadership experience as a former sports captain and competitive athlete. Whether training for marathons or diving into the latest AI research, I approach every challenge with analytical thinking and determination.
I thrive at the intersection of AI, data engineering, and product development, transforming complex problems into scalable solutions.
Aug 29 to Sep 28, 2026
Most of this month's commits went to private repos. In public, the work was Coo: I split the kernel into an always-on edge and a Mac-side node that resolves and executes requests, then made the edge run on a Linux VPS with a systemd installer, tailnet address detection on Linux and a role-aware LaunchAgent installer, and wrote up the split and the deployment. Settl got a few pushes but no new commits.
My projects
- abhinavaditya811/coo: 11 commits · Document the split and the VPS deployment
- abhinavaditya811/settl: 3 pushes
Private work: 50 commits in private repos
Languages: Python
All time: contributions to 10 repositories I don't own, including HKUDS/nanobot, asreview/asreview, caipe-io/ai-platform-engineering.
Updated automatically, last run Sep 28, 2026
- Email: aditya.ab@northeastern.edu
- LinkedIn: linkedin.com/in/abhinavaditya811/
- Portfolio: 1abhinavaditya.vercel.app
