I'm a quantitative researcher and systems builder working at the intersection of high-performance computing, capital markets, and machine learning. My focus is on what actually moves the needle at low-latency scale - cache behaviour of financial kernels, foundation models trained on market data, and the tooling layer around modern research workflows.

I'm drawn to problems where the maths, the hardware, and the market all have a vote. I build a lot. I read a lot. I ship the things I read about.

Currently a researcher at Indiana University - Luddy School of Informatics, Computing and Engineering, focused on cache-aware numerical methods for quantitative finance.

timeline
    title From curiosity to shipping - 2019 → 2026
    2019 : Linear Algebra foundations (MIT 18.06 / EE16A)
         : First taste of numerical thinking
    2025 Q4 : Paul Wilmott quant finance practice
            : Generative ML experiments (private)
            : First hackathon-grade shipped UI (skillconnect-jk)
            : Image recognition + captioning prototype
    2026 Q1 : LABLAB hackathons (AI + startup tracks)
            : NCAA bracket analytics (Kory_The_Cat-NCAA)
            : Robot simulation environment
            : CPS Digital Design (IU E222) - embedded + FPGA
    2026 Q1–Q2 : Research thread opens - cache heuristics in finance
               : Claude Hackathon - WHO-aligned telehealth
               : Clap / OpsPilot - agentic workflow automation
               : Personal site rebuild (Next.js)
               : AI Ethics Coach - privacy-first ChatGPT companion
    2026 Q2 : finance-cache-hpc - EPYC + PAPI study (28× · 1,657× · 500×)
            : ResumeForge - local-first explainable resume AI
            : entrepreneur-persona - pitch coach for MIT/Harvard/Kelley comps
            : DeskFlow Native - cross-platform workflow manager
            : Abstraction Dictionary - book project kickoff
            : Map_Projects_MAC - portfolio index goes live

The through-line: learn by building, measure before tuning, ship the thing. Every entry on that line is a public (or private-on-request) repo you can open today.

• finance-cache-hpc - L1 cache characterisation of four quant kernels on AMD EPYC using PAPI
• Research_HPC_QFinance_Cache - the broader research thread this sits inside
• QuantumMCL-Spring26 - (private) heuristic cache mechanisms for finance workflows
• ResumeForge - local-first, explainable resume tailoring with LaTeX + Overleaf sync

_{🔒 = private, available on request · full index → Map_Projects_MAC}

Repos I've cloned to dissect, not to fork. Honest credit goes to the upstream authors:

• HPC / quant - NVIDIA/cutlass · siboehm/SGEMM_CUDA · KxSystems/kdb-taq
• ML systems - karpathy/arxiv-sanity-lite · openai/parameter-golf · Zjh-819/LLMDataHub
• Agents / personal tooling - AmitSubhash/autolog · msitarzewski/agency-agents
• Coursework - Developer-Y/cs-video-courses · Strang EE16A notebooks · IBM Watson NLP template

Full list with annotations → Following.md

Curiosity is the unit. Everything else - papers read, kernels tuned, repos shipped - is just accumulated interest on it.

I think the best researchers are the ones who can also build, and the best builders are the ones who read the papers. I try to be useful on both sides of that line.