╔══════════════════════════════════════════════════════════════╗
║           ANURAG S. CHATTERJEE  //  MS Computer Engineering  ║
║                    Columbia University, NYC                   ║
╚══════════════════════════════════════════════════════════════╝

I'm an Electrical & Computer Engineer with expertise in systems, IoT, AI, and applied engineering. I enjoy working at the intersection of hardware, software, and large-scale systems, with a passion for systems engineering, robotics, AI, mathematics, finance, and quantitative problem-solving.

I build things that run on real hardware, write code that solves real problems, and document everything carefully enough that someone else can pick it up and continue. Previously worked at Schlumberger and Thales and multiple deep-tech startups in Singapore.

Working on deep learning for electromechanical wave imaging (EWI) of the heart — using ultrasound-based techniques to non-invasively monitor the electrical impulses that propagate through cardiac tissue during contraction and arrhythmia. The goal is to map electromechanical activation sequences to help diagnose and understand conditions like atrial fibrillation, ventricular tachycardia, and other cardiac arrhythmias.

Deep Learning Cardiac Imaging Ultrasound Signal Processing Electromechanics Biomedical Engineering

Implemented semidefinite programming and convex optimization in C on Raspberry Pi 5 for sustainable microgrid energy management in data centres, with hardware-in-the-loop testing via MATLAB and Simulink.

→ Full project repository

Python C MATLAB CVXPY CVXPYgen Clarabel Raspberry Pi 5 TCP/IP

All stats below are always visible — no external images or API calls required.

Language / Tool	Used In
Python	Convex Optimization · Computer Vision · IoT · NLP Benchmarking
C	SDP Energy Management · Embedded Solvers · CVXPYgen
C++	GPU DVFS · HPC Sorting · Arduino Signal Visualization
MATLAB	Microgrid Optimization · Test Automation · HIL Testing
CUDA	GPU Microarchitecture · DVFS Policies · LLM Training
Verilog	FPGA MLP Accelerator · Xilinx Zynq · Hardware Inference
PyTorch / TensorFlow / Keras	Deep Learning · Cardiac EWI · Computer Vision
Embedded C	Raspberry Pi · ESP32 · ARM · Sensor Systems
SQL	IoT Deception Detection · Data Pipelines

⚡ SDP Energy Management in C on Raspberry Pi Multi-period DC microgrid OPF via SDP relaxation compiled to C with CVXPYgen. 5-bus network, 24-hour horizon, 845 variables, 1667 constraints, 24× (5×5) PSD cones. 304 fixed permutations — 100% pass rate. Python C MATLAB CVXPY Raspberry Pi 5 TCP/IP

🖥️ Optimal Sorting for 4GB LLM Datasets MergeSort + OpenMP across 1B+ 32-bit integers. 6.2× speedup on 16-core systems via cache-aware partitioning and NUMA-optimized memory allocation. C++ OpenMP NUMA Computer Architecture

⚙️ DVFS for GPU Microarchitecture in Data Centres DVFS on NVIDIA Ampere and Volta GPUs. 18% reduction in energy-delay product during LLM training. C++ CUDA GPU Architecture

🦯 NAVI — AIoT Smart Blind Cane LiDAR + GPS + motor control on ESP32 (I2C, PWM). Real-time obstacle detection + cloud monitoring. 90% accuracy over 0.5–1m. Audience Choice Award, Fall 2025. ESP32 LiDAR I2C PWM AWS IoT

🧠 MLP Neural Network on FPGA 100-layer MLP accelerator in Verilog on Xilinx Zynq FPGA. 120 fps with INT8 quantization — 94% accuracy vs FP32. Verilog FPGA Xilinx Zynq Hardware Acceleration

👷 Multi-Object Tracking — PPE Compliance YOLOv5 + DeepSORT on 4K/30fps. 81% mAP@0.5 (from 65%). 500K+ frames/day, sub-100ms latency, 12 sites. YOLOv5 DeepSORT FastAPI Computer Vision

🔍 IoT Deception Detection for Examinations Arduino sensor array (GSR, heart rate, temperature) + Random Forest / SVM. 90% accuracy across 200+ subjects. Arduino ML SQL Sensor Fusion

🤝 Columbia-Amazon Bedrock Innovation Challenge 2025 Multi-agent AI system using Amazon Bedrock (Claude 3.5) for FMCG supply chain optimization — LLM demand forecasting, inventory management, and logistics routing. Amazon Bedrock Claude 3.5 Multi-Agent AI LLM

➡️ See pinned repositories below for full documentation, setup instructions, and results.