Advaith Vecham CS student at the University of Washington building systems software, LLM agent architectures, and low-level performance work. I'm a Computer Science student at the University of Washington (Seattle), expected 2027. I work across the stack — from Rust contributions to the Linux kernel's timing subsystem, to CUDA kernels for quantized LLM inference, to autonomous browser agents that plan and self-heal. Most recently I've been building Insuo, a hierarchical plan-first browser agent that also learns and creates automations without prompting, and contributing to the Rust for Linux subsystem. I was also selected for YC Startup School (top 10% of 30,000+ applicants). When I'm not deep in kernel code or CUDA kernels, I'm usually doing competitive math/programming (AIME, USACO Gold) or exploring new corners of systems and ML. WHAT I WORK ON - contribute Rust abstractions to the Linux kernel. - write CUDA kernels for quantized LLM inference. - build autonomous agent architectures (planning, self-healing, context management). - compete in math and programming (Codeforces, PUTNAM). SKILLS C++, CUDA, LLM, Python, React, Rust, Agents, Desktop, Docker, Emulation, Fly.io, Gemini API, Git, Infra, ML, ML Infra, Node.js, Performance, PyTorch, Quantization, Systems, Tauri, TensorRT-LLM, TypeScript, Vite, WebKit PROJECTS Insuo — AI-Powered Autonomous Browser Agent Stack: Python, TypeScript, C++, WebKit, Gemini API, Tauri Tags: Agents, LLM, Desktop A hierarchical plan-first agent (Goal → Parts → Steps) that decomposes high-level tasks into verifiable sub-goals, cutting redundant LLM calls by 60% vs. pure ReAct by executing deterministic steps open-loop. Self-healing loop catches mistakes and retries with better insight. Compiled to a downloadable desktop app via Tauri. Mixed-Precision Quantized GEMM Kernel Stack: C++, CUDA, Git Tags: CUDA, Performance, ML Infra A high-performance C++/CUDA linear layer kernel fusing INT4/INT8 weight de-quantization with FP16 tensor core matmul to accelerate LLM inference. Cut L1 cache misses 50% by packing low-precision weights into integer types with fast bit-shift unpacking, and eliminated shared memory bank conflicts via memory padding. CHIP-8 Emulator for macOS Stack: Rust Tags: Emulation, Systems Source: https://github.com/adsdemaybe/chip8_emulator A full-stack hardware emulator replicating the CHIP-8 interpreter: 4KB memory layout, 16 general-purpose registers, hardware stack, and a fetch-decode-execute cycle handling all 35 opcodes. 60Hz pixels crate based rendering pipeline. Supports Superchip, CHIP-8 and Atari game ROMs, and tested with Space Invaders, and various other public IP games. Can set clockspeeds and FPS dynamically during build. First time making something in Rust, and actually learn how to use macros, defining enums, shortcuts while coding, etc. CHIP-8 build was hardcoded, Superchip implementation was then added later with the help of AI. Finetuned Qwen 3.5 Browser Agent Stack: Python, PyTorch, TensorRT-LLM, Docker Tags: ML, LLM, Quantization Finetuned a Qwen 3.5 9B model for autonomous web navigation with specialized tool-calling loops over real-time financial market environments. Deployed via a high-performance serving engine with optimized KV-cache management, and quantized to 4-bit (AWQ/GPTQ) for over 60% memory reduction at high-density edge deployment. Terminal Portfolio Stack: React, Vite, Rust, Node.js, Fly.io Tags: React, Infra Source: https://github.com/adsdemaybe/portfolio_website This site — a portfolio that behaves like a terminal by default, with a fixed set of page commands falling through to a real backend Linux sandbox (per-session Firecracker VMs, real gcc/rustc/python3, encrypted snapshot save/restore) for anything else, plus a modern "Easy Mode" hiding underneath. CONTACT Email: advaiv2@cs.washington.edu GitHub: https://github.com/adsdemaybe LinkedIn: https://www.linkedin.com/in/advaithvecham/ Location: Redmond, WA