I build and debug real systems across compilers, robotics, and edge AI — where failures are subtle, performance matters, and correctness is non-negotiable.

This is not a collection of demos. This is a portfolio of engineering depth.

• 🧠 Compiler & Toolchain Engineering (MetaWare, low-level debugging)
• ⚙️ Systems Programming (C/C++, runtime behavior, memory-level analysis)
• 🤖 Robotics & Perception (ROS2, SLAM, real-time pipelines)
• 👁️ Edge AI Systems (YOLO, Raspberry Pi, deployment constraints)
• 🧪 Automation & Qualification (reproducibility, CI pipelines, validation)

Most engineers:

• write code
• train models
• build features

I:

• debug compiler-generated assembly
• validate runtime behavior under safety constraints
• build systems that fail → get traced → get fixed properly

• Multi-version compiler validation (MetaWare)
• Automated pipeline: compile → execute → validate → report
• Known defect reproduction (broken vs fixed comparison)
• Assembly-level debugging using TRACE32
• MISRA / ISO 26262 aligned validation

👉 I don’t just use compilers — I verify and break them.

• Simulated sensors + depth estimation
• YOLO object detection integration
• Occupancy grid mapping + SLAM
• RViz visualization + control loop

👉 Not just perception — full system integration.

• YOLO-based detection (~98% accuracy)
• Raspberry Pi 5 + Tapo camera integration
• Flask dashboard + cloud deployment
• Latency and accuracy optimization

👉 Built for real-world constraints, not ideal conditions.

• 150+ unit tests (~95% coverage)
• Black-box + system-level validation
• CI-friendly pipelines (CMake + Python)
• Reproducible defect workflows

👉 If it can’t be tested properly, it’s not engineering.

Languages C / C++ / Python

Systems & Tools Linux • CMake • TRACE32 • nSIM Jenkins • GitHub Actions

Domains Compiler Toolchains • Embedded Systems ROS2 • SLAM • Computer Vision

• Failure → trace to root cause
• Unclear behavior → inspect assembly
• Non-reproducible → fix the system, not the symptom

[Source Code]
      ↓
[Compiler (Multiple Versions)]
      ↓
[Binary Execution on Target]
      ↓
[Output Capture]
      ↓
[Validation Engine]
      ↓
[Report + Defect Tracking]

[Sensors / Simulation]
        ↓
[Depth Estimation]
        ↓
[YOLO Detection]
        ↓
[Occupancy Grid / SLAM]
        ↓
[Decision Logic]
        ↓
[Control Output]

[Camera Input]
      ↓
[Preprocessing]
      ↓
[YOLO Model]
      ↓
[Detection Results]
      ↓
[Backend API]
      ↓
[Dashboard / Cloud]

1. Compiler Qualification Framework
2. Known Defects Repository
3. ROS2 Perception System
4. Occupancy Monitoring System
5. Automation Testing Framework
6. Low-Level / FPGA / Experimental Work

• Demo-only projects
• Black-box dependency reliance
• “Works on my machine” engineering

• Compiler Engineering
• Systems Programming
• Embedded & Low-Level Development
• Robotics Systems Engineering

• GitHub: https://github.com/K-Ajay-Kumar
• LinkedIn: https://linkedin.com/in/ajay-kumar-k-3b53821b4

If you're working on compilers, low-level systems, or real-time pipelines — we’ll have something meaningful to talk about.