Skip to content
View suntosh's full-sized avatar
🏠
Working from home
🏠
Working from home
  • Purdue University
  • Richardson, Texas
  • 23:30 (UTC -12:00)

Highlights

  • Pro

Block or report suntosh

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
suntosh/README.md

Santosh Ahuja

Engineering executive and architect. Twenty years building systems and the organizations that ship them in payments, healthcare, enterprise SaaS, defense, and AI infrastructure.

📍 Richardson, TX  ·  🎓 MS ECE, Purdue


🧭 What I do

I operate at both ends of the stack and treat that as one job, not two.

⬇️ Down the stack — RF and VLSI (5.25 GHz LNA, 6G receiver in 45nm CMOS), compute-in-memory macro characterization, CUDA kernels (GEMM, Tensor Core WMMA, Flash Attention), embedded systems and RTOS, CAN/CANOpen, ROS2.

⬆️ Up the stack — carrier-grade platform architecture, IoT gateways, SIP/VoIP, data center networks, LLM routing and orchestration, semantic layers over fragmented enterprise estates.

↔️ Across the org — CTO, VP Engineering, Director, Chief Architect. Built engineering organizations from 3 to 60+ across defense, telecom, healthcare, and fintech.


💻 Languages

C++ C Python Java C# Go TypeScript CUDA Verilog Bash


🔬 Silicon & EDA

Cadence Virtuoso Spectre Monte Carlo ADS GPGPU-Sim

🧪 Where I use them — 6T SRAM compute-in-memory macro characterization (dual-Vmin at 0.55–0.65V, mixed-Vt HVT-storage / LVT-access, 58% energy reduction at 0.65V) · 5.25 GHz LNA · 45nm CMOS 6G receiver front-end · RF conformance across GPS L1/L5, WiFi, and BLE · high-speed mixed-signal design · IC/MEMS fabrication.


🛠️ Stacks I build on

🤖 ML & accelerators

PyTorch NVIDIA Hugging Face NumPy

🏗️ Platform & data

Kafka Redis PostgreSQL Spring Boot .NET

☁️ Infrastructure

Azure Docker Linux Git

🦾 Embedded & edge

ROS2 RTOS CAN CMake


The unifying thread: I care about what the abstraction is hiding. Usually it's a transistor.

Pinned Loading

  1. Generative-ai-math Generative-ai-math Public

    Capture of Math notes around gen ai - captured from NPTEL course.

  2. MLEngineer-Reference MLEngineer-Reference Public

    An ML Engineer's reference - a toolbox for production ops

    TeX

  3. Go-Specialization Go-Specialization Public

    Solutions to 3-Course Go specialization series from 'University of Irvine' delivered through Coursera

    Go

  4. SDN-Simulation SDN-Simulation Public

    SDN control plane over UDP. Dijkstra routing, keep-alive failure detection, and sub-6s reconvergence after switch or link loss.

    Python 1

  5. minnet minnet Public

    Dependency-free C++17 neural network library. Every gradient verified against finite differences (117 tests)

    C++

  6. ece608-toposort ece608-toposort Public

    topological-sort, graph-algorithms, dag, algorithms, python, kahns-algorithm

    Python