About Research Projects Skills Experience Contact

Hello, I'm

Raj Mali

I design |

Bridging biological neural architectures with silicon — from transistor-level SRAM to spiking neural networks on FPGA.

0 CGPA / 10
0 Power Reduction
0 RISC-V Processor

Scroll to explore

About Me

I'm a third-year Electronics & Computer Science student at Vidyalankar Institute of Technology, Mumbai, with a deep passion for understanding computing at its most fundamental level — the transistor.

My journey started with building circuits from raw logic gates, deliberately avoiding microcontrollers to truly understand digital design. That instinct to reach for the transistor before the library call now drives my research in neuromorphic computing and VLSI design.

I believe the next generation of computing will look less like a CPU and more like a brain. My work sits at that intersection — designing hardware that processes information the way biological systems do: efficiently, adaptively, and at ultra-low power.

Education

B.Tech, ECE
VIT Mumbai | 9.44 CGPA
Gold Medal in CS (10th)

Research Focus

Neuromorphic Computing
SRAM Design & Analysis
Spiking Neural Networks

Achievements

NPTEL Silver Medal
Hackathon Winner
District Badminton Champ

Research

Paper

Voltage Scaling and Stability Trade-offs in 6T versus 12T SRAM Cells

Harshal More*, Sanika Sonavane*, Raj Mali*, Prof. Akhil Masurkar

Vidyalankar Institute of Technology · 180nm CMOS · Cadence Virtuoso

Power Reduction
2.6× Better SNM
0.3V Min Operating Voltage
4.82× Area Trade-off

Performed butterfly curve analysis and DC sweep characterization comparing 6T and 12T SRAM topologies. Demonstrated that 12T cells eliminate read disturb entirely and enable sub-0.5V operation with 9× dynamic power reduction — at the cost of 4.82× area overhead. Directly relevant to ultra-low-power IoT, biomedical SoCs, and neuromorphic memory arrays.

SRAM Voltage Scaling SNM Butterfly Curve Low-Power IoT
Active Research

Neuromorphic Computing & Brain-Inspired Architectures

Exploring the design space of hardware-efficient brain-inspired computing. Currently studying Artificial Neural Networks (ANN) and Spiking Neural Networks (SNN) with a focus on translating biological neuron dynamics — synaptic plasticity, membrane potentials, spike-timing-dependent plasticity (STDP) — into scalable CMOS and FPGA implementations.

  • Spike generator architectures for SNN on Xilinx FPGA using stochastic neuron models
  • Floating-gate MOSFET neuron concept: multiple control gates as dendritic inputs
  • Sub-threshold MOSFET operation for analog synaptic weight emulation

Projects

VLSI

32-bit RISC-V Processor

Designed a complete 32-bit RISC-V processor with 16-bit ALU, external memory interfacing, and multi-sensor peripheral integration (DHT11, ultrasonic, IR) on Xilinx FPGA.

VerilogVivadoFPGAFSM
Research

SNN Spike Generator on FPGA

Hardware spike generator emulating biological neuron spiking behavior using stochastic RNG-based neuron models. Targeting minimal power and resource utilization for edge neuromorphic applications.

VerilogVivadoSNNArtix-7
IP Core

UART-BLE Communication IP

Reusable IP core for UART-to-BLE data bridging. Deployed in a smart-home prototype enabling real-time mobile-to-FPGA control of lights and motors via Bluetooth.

VerilogIP DesignBLEFPGA
Competition

E-Yantra Maze Solver

IIT Bombay's E-Yantra robotics competition. Designed a maze-solving algorithm implemented entirely in Verilog with FSM-controlled multi-sensor integration.

VerilogFSMSensors
VLSI

Capacitive Touch Piano

Touch-sensitive musical instrument on FPGA using frequency modulation and scaling techniques for accurate multi-octave tone synthesis.

VerilogFPGAFreq. Modulation
Software

GhostType — AI Typing Assistant

Multi-process Electron + Python desktop app integrating Mistral AI & Gemini Vision API. Features multi-process IPC, thread-safe keystroke monitoring, voice-to-text, and <200ms latency. 42 unit tests.

ElectronPythonMistral AIGemini
Software

ELECTRO-Inventory

Full-stack real-time inventory management system with SQL backend, dual-login portal, low-stock alerts, and profit tracking. Deployed in multiple retail stores.

SQLHTML/CSS/JSFull-Stack
Hardware

Pure Logic Digital Systems

Fastest Finger First game, Ultrasonic Radar (30m range), Peltier liquid cooler with live temp readout — all built with discrete gates and flip-flops, zero microcontrollers.

AND/OR GatesD Flip-FlopsPCB

Technical Skills

HDL & VLSI Design

VerilogVHDLFPGA DesignIP DevelopmentCMOS DesignSRAM ArchitectureFSM DesignTestbench & Verification

EDA Tools

Xilinx VivadoCadence VirtuosoIntel QuartusMATLAB / SimulinkEasyEDA

Processor & Architecture

RISC-VALU DesignMemory InterfacingPeripheral IntegrationXADC

Programming

PythonC / C++JavaSQLJavaScriptMATLABDart (Flutter)CUDA (learning)

Tools & Platforms

Git / GitHubElectronDjangoReactLinux

Hardware

PCB Design & SolderingOscilloscopeLogic AnalyzersIoT PrototypingSignal Analysis

Experience

R&D Intern

May 2025 — Aug 2025

Prosmart Solutions, India

  • Prototyped hardware products (dehumidifier, automated alert systems) using discrete logic components for high reliability and low cost
  • Developed a Flutter mobile app for inventory management with Google Sheets API integration
  • Performed component-level debugging, soldering, and repair of industrial PCBs

B.Tech — Electronics & Computer Science

2023 — 2027

Vidyalankar Institute of Technology, Mumbai

  • CGPA: 9.44/10 (average) · 9.82/10 (highest semester)
  • IEEE Photography & Videography Head · VCC Senior Core
  • Represented college in inter-college chess and badminton tournaments

Achievements

🏅

NPTEL Silver Medal

Joy of Computing using Python — IIT Madras (81%)

🏆

1st Prize

DigiGo Hardware Hackathon

🏈

3rd Prize

Zero-True International Hackathon

🏅

District Champion

Badminton · 4th in Intra-State

🏏

National Cricket

Under-13 National Level Representative

🥇

Gold Medal

Computer Science — 10th Standard (ICSE)

Certifications

Digital Design with Verilog
CMOS Digital VLSI Design
NPTEL — Joy of Computing (Python), IIT Madras
MATLAB Onramp & Simulink Onramp — MathWorks
Control Systems Onramp — MathWorks

Let's Connect

Open to research collaborations, internship opportunities, and conversations about neuromorphic computing and VLSI design.