Skip to content
View Varun2459's full-sized avatar

Block or report Varun2459

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
Varun2459/README.md

Hi 👋, I'm Varun Molleti

FPGA & Digital Design Engineer — RTL · HLS · VLSI | MSc Telecommunications, University of Liverpool


About Me

  • 🔭 Dissertation: Building a real-time FPGA image processing accelerator (Canny-style pipeline) using HLS C++ on Intel Cyclone V — achieving II = 1, 100 Mpixels/s throughput, and timing closure at Fmax ≥ 138.87 MHz
  • 🎓 MSc Telecommunications & Wireless Systems at University of Liverpool — Distinction predicted
  • B.Tech highlight: Designed energy-efficient approximate multipliers in Verilog (Cadence Genus, 90 nm CMOS) — 96.9% power reduction and 95.5% area reduction vs. standard array multipliers
  • 🛠️ Proficient in Verilog, VHDL, SystemVerilog, HLS C++ (Intel i++), with hands-on EDA experience across Quartus Prime, Platform Designer (Qsys), QuestaSim, and Cadence Genus/Tempus
  • 📫 Reach me at varunvenkatatej7@gmail.com | Open to UK graduate FPGA, VLSI, and Digital Design roles

Technical Skills

HDL & Hardware Description

Verilog VHDL SystemVerilog HLS C++

EDA & Design Tools

Quartus Prime Cadence Genus QuestaSim Platform Designer Cadence Tempus

Programming & Scripting

C++ Python C MATLAB Git

ML & Data

scikit-learn OpenCV pandas NumPy


Featured Projects

Real-Time FPGA Image Processing Accelerator (MSc Dissertation)

Full Canny-style edge detection pipeline (Gaussian blur → Sobel → NMS → double-threshold) implemented as a single HLS C++ kernel on Intel Cyclone V with Avalon-ST interfaces and live VGA output.

  • Result: II = 1 | 100 Mpixels/s (~325 fps @ 640×480) | Fmax ≥ 138.87 MHz | 7,398 ALMs (23%)
  • Tools: Intel HLS Compiler (i++), Quartus Prime 18.1, Platform Designer (Qsys), DE1-SoC, SDC constraints

Energy-Efficient Approximate Multiplier for DSP Applications (B.Tech Final Year)

Designed novel Type-1 and Type-2 approximate compressors in Verilog to eliminate carry-chain logic; validated output quality using PSNR/SSIM on an image sharpening pipeline.

  • Result: 96.9% power reduction | 95.5% area reduction | 92.3% delay reduction (vs. standard 8-bit multiplier)
  • Tools: Verilog, QuestaSim 10.6c, Cadence Genus (90 nm CMOS), Cadence Tempus, MATLAB

ASL Hand Gesture Recognition (ML System)

Vision-based real-time sign language classifier — trained and benchmarked Random Forest, SVM, and KNN on 747-sample hand landmark data; deployed live via OpenCV camera pipeline.

  • Result: Random Forest selected as primary classifier based on F1 score and noise robustness
  • Tools: Python, scikit-learn, OpenCV, pandas, NumPy

GitHub Stats

  


Connect

   


Open to UK graduate roles in FPGA Design · VLSI · Digital Hardware · Semiconductor Engineering

Popular repositories Loading

  1. Varun2459 Varun2459 Public

  2. FPGA-Image-Processing-HLS FPGA-Image-Processing-HLS Public

    Real-time Canny-style edge detection pipeline on Intel Cyclone V - HLS C++ kernel, II=1, 100M pixels/s, Fmax ≥ 138.87 MHz, live VGA output

    C++

  3. Approximate-Multiplier-90nm-CMOS Approximate-Multiplier-90nm-CMOS Public

    ASIC synthesis of approximate constant-carry multipliers at 90 nm CMOS. Cadence Genus/Tempus, QuestaSim, MATLAB PSNR/SSIM validation. 96.9% power reduction, 95.5% area reduction vs standard array m…

    Verilog