School of Electronics Engineering, Vellore Institute of Technology (VIT)
I am an Electronics Engineer specializing in the intersection of hardware architectures, real-time embedded systems, and on-device machine learning (Edge AI / TinyML). My core engineering focus spans:
- β‘ Embedded Systems & Firmware Development: Designing deterministic bare-metal and RTOS firmware for ARM Cortex-M and ESP32 microcontrollers, managing standard bus protocols (I2C, SPI, UART, CAN).
- ποΈ Digital System Design & FPGA/RTL: Authoring modular, synthesizable Verilog HDL architectures, parameterizable clock dividers, and finite state machines (FSMs) verified through ModelSim/QuestaSim.
- π Control Systems & Power Electronics: Engineering adaptive, safety-critical control algorithms (Parallel Online Recursive Least Squares, Constrained Model Predictive Control) for multi-cell Lithium-ion Battery Management Systems (BMS).
- π§ Edge AI & TinyML: Quantizing, optimizing, and deploying deep learning models (e.g., autoencoders and convolutional neural networks) directly onto memory-constrained microcontrollers for real-time anomaly detection and edge inferencing.
Microcontrollers : ESP32, STM32 (ARM Cortex-M), Arduino, TI MSP430
RTOS & Firmware : FreeRTOS, Bare-Metal C/C++, Embedded C, PlatformIO
Hardware Protocols : I2C, SPI, UART, CAN, PWM, GPIO, ADC/DAC
Sensors & Actuators : MPU6050 (IMU), Current Shunts, Thermal Sensors, Motor Drivers
Hardware Languages : Verilog HDL, RTL Design, SystemVerilog (Verification)
FSM & Digital Logic : Moore & Mealy State Machines, Clock Domain Crossings, Clock Dividers
FPGA Platforms : Intel Cyclone IV, Xilinx Spartan-6, Quartus Prime, Xilinx Vivado
EDA & Simulation : ModelSim, QuestaSim, Icarus Verilog, GTKWave, NI Multisim, LTspice
Control Algorithms : Recursive Least Squares (RLS), Model Predictive Control (MPC), State-Space
Modeling & Tools : MATLAB, Simulink, Simscape Electrical, Python Control Systems
Digital Comms : BPSK, QAM, AWGN Channel Modeling, Monte Carlo BER Analysis, DSP
Edge AI & TinyML : TensorFlow Lite for Microcontrollers (TFLM), Edge Impulse, TinyML Autoencoders
Machine Learning : TensorFlow, Keras, PyTorch, Scikit-learn, OpenCV
Architectures : Convolutional Neural Networks (CNNs), MobileNetV2, Inverted Residuals, Autoencoders
Deployment : Model Quantization (INT8/Float16), Pruning, Streamlit Edge Dashboards
Version Control : Git, GitHub CLI (gh), CI/CD Workflows
Software & Web : Python (NumPy, SciPy, Matplotlib), Next.js, React, TypeScript, C/C++, WebExtensions (Manifest V3)
Operating Systems : Linux (Ubuntu), Windows, Embedded Linux
| Project | Domain | Architecture & Highlights |
|---|---|---|
| SmartCharge-Adapt | BMS & Adaptive Control | Per-cell parameter-adaptive fast-charging control for degrading series Li-ion packs using parallel online RLS and constrained MPC. Enforces an absolute 4.200V ceiling per cell with 0.0 mV lifetime overshoot across 1,000 aging cycles. |
| Traffic-Light-Controller | FPGA & Verilog HDL | Synthesizable two-way digital traffic light controller featuring asynchronous pedestrian request latching, safety-critical emergency preemption, and parameterizable dual-mode clock division. |
| Digital Image Transmission over AWGN | Digital Communications | Performance analysis and Monte Carlo simulation of BPSK and 16-QAM constellations over AWGN channels with Bit Error Rate (BER) and Peak Signal-to-Noise Ratio (PSNR) validation in MATLAB. |
| Project | Domain | Architecture & Highlights |
|---|---|---|
| VibroShield-Edge-SHM | Edge AI & IoT | Autonomous Edge Structural Health Monitoring system running TinyML Autoencoders directly on an ESP32 to detect mechanical anomalies from IMU vibration telemetry, synchronized with a real-time Streamlit dashboard. |
| Flower-Recognition | Deep Learning / CV | Empirical benchmark comparing a custom scratch-built CNN against MobileNetV2 transfer learning across 5 botanical species, evaluating training convergence, augmentation, and out-of-distribution web inference. |
| DevForge (HackerStreet Rebels) | AWS Hackathon | Web IDE extension enforcing architectural discipline through continuous AST constraint validation, real-time risk scoring, and AWS Cost Whisperer cloud expenditure governance. |
| Project | Domain | Architecture & Highlights |
|---|---|---|
| VTOP-Vault | Browser Extension (Manifest V3) | Lightweight, automated Manifest V3 browser extension for VIT students. Intercepts CSRF/Auth-scoped POST streams to batch-download course materials into structured local hierarchies (VIT Downloads/Course/Faculty/Module/filename.pdf) with custom "Bully Blocker" collision avoidance. |
- π§ Email: reachfortheskiez@gmail.com
- πΌ GitHub: @Sky-Code06
- π« Institution: School of Electronics Engineering, Vellore Institute of Technology (VIT)