Skip to content
View chinmaymahananda's full-sized avatar

Block or report chinmaymahananda

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 donโ€™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 userโ€™s behavior. Learn more about reporting abuse.

Report abuse
chinmaymahananda/README.md

Hi, I'm Chinmay ๐Ÿ‘‹

Hardware & Silicon Engineer โ€” VLSI/Mixed-Signal Design ยท Digital & PCB Hardware ยท ML-Hardware Systems M.S. ECE @ Northeastern University (Hardware & Software for Machine Intelligence track, GPA 3.92/4.0) ยท Published 6G antenna researcher ยท Building across the stack from transistors to Verilog to ML accelerators.

๐Ÿ“ Seattle, WA ย |ย  ๐ŸŽฏ Seeking Hardware/Silicon Engineering Internships & New-Grad Roles ย |ย  โœ‰๏ธ mahananda.c@northeastern.edu ย |ย  ๐Ÿ”— LinkedIn


โš™๏ธ What I work on

I design and verify hardware end-to-end โ€” RTL, CMOS/digital circuits, PCBs โ€” and build the classical & deep learning systems that run on top of them.

  • VLSI / Circuit Design: CMOS & digital logic, schematic capture & simulation (Cadence Virtuoso, Spectre), MOSFET Iโ€“V characterization, DC/transient analysis, logical-effort sizing
  • Digital Design / HDL: Verilog, RTL-to-GDSII ASIC flows (Yosys, OpenROAD, Sky130), parameterizable datapath design, self-checking testbenches
  • PCB / Embedded: Schematic capture, layout & fabrication (KiCad, Altium), board bring-up & fault isolation, Embedded C
  • ML / Computer Vision: PyTorch, CNN-based detection & embedding models (YOLO11, EfficientNet, transfer learning), classical ML for imbalanced classification (XGBoost, scikit-learn)

๐Ÿ”ง Tools & Languages

Verilog Python Embedded C MATLAB SPICE Cadence Virtuoso Spectre LTspice CST Studio Suite KiCad Altium PyTorch scikit-learn XGBoost Git Icarus Verilog Logisim

๐Ÿš€ Featured Projects

Project Description
riscv-pipelined-core 4-stage pipelined RV32I core (IFโ†’EXโ†’MEMโ†’WB) with hazard forwarding, load-use stalling, and branch handling, verified bit-exact against a from-scratch golden model โ€” carried through a full open-source ASIC flow (Yosys, OpenROAD, Sky130 PDK) to signed-off GDSII at 333MHz, zero DRC/LVS violations. A 500MHz stretch attempt was honestly documented as a failed timing closure rather than discarded.
systolic-mac-array Parameterizable output-stationary systolic MAC array built bottom-up from a single MAC PE, extended into a full INT8 CNN inference accelerator (Conv1โ†’Conv2โ†’FC) with interlayer requantization. Traced a silent numerical drift on non-square input tiles to an edge case in the accumulator's saturation logic via waveform-level debugging, patched it, and re-verified bit-exact against a Python golden model across all 20 calibration samples.
8-Bit ALU & 4:1 MUX 8-bit ALU built from a 1-bit cell in Logisim (AND/OR/XOR/NOR, two's-complement add/subtract, SLT); 4:1 MUX built and transient-verified in Cadence Virtuoso/Spectre across all four select states.
PCB Fabrication & Bring-Up End-to-end board ownership โ€” schematic capture, layout, fabrication, assembly, and functional test, bringing up power and signal nets on first power-on. Isolated a failed signal net to an improperly routed ground return via systematic continuity/probe testing rather than reflow guesswork.
Pokรฉmon TCG Card Recognition Team project (with Jeffrey Ma, Rayne Liu) โ€” YOLO11-Nano OBB detector + EfficientNet-B0 triplet-loss embedding pipeline for sub-2-second recognition across 558 card embeddings, trained on a single RTX 3050 with augmentation to close the sim-to-real gap.
Credit Card Fraud Detection Benchmarked 5 ML models on 284,807 transactions at 0.17% fraud prevalence โ€” ~0.99 AUC-ROC via leak-free SMOTE and recall-prioritized threshold tuning under asymmetric misclassification cost.

๐Ÿ“„ Publication

"Design of 6G Antenna" โ€” IJNRD 2024, Vol. 9 Issue 6 (ISSN 2456-4184), first author โ€” millimeter- and micrometer-scale microstrip patch antennas for 6G THz-band communication, designed and simulated in CST Studio Suite. Read it โ†’

๐ŸŽ“ Education

M.S. Electrical & Computer Engineering โ€” Northeastern University (Seattle), GPA 3.92/4.0 โ€” Sep 2025โ€“Present

B.E. Electronics & Communication Engineering โ€” Dr. Ambedkar Institute of Technology, Bengaluru โ€” 2020โ€“2024

Recipient of the Seattle Campus Scholarship and International Impact Award โ€” competitive merit scholarships covering 30% of program tuition combined.

๐Ÿ“œ Certification

Emertxe Certified Embedded Professional (ECEP) โ€” Emertxe Information Technologies, 2025 โ€” Linux Systems, Advanced C, Data Structures, Linux Internals & Networking, Micro Controllers, C++ Programming. Verify โ†’


โœ‰๏ธ Reach me at mahananda.c@northeastern.edu or on LinkedIn

Pinned Loading

  1. riscv-pipelined-core riscv-pipelined-core Public

    4-stage pipelined RV32I RISC-V core (IF->EX->MEM->WB) with hazard forwarding and branch handling -- full RTL-to-GDSII tapeout via OpenLane/Sky130, timing closed at 333MHz, 0 DRC/LVS violations.

    Verilog 1

  2. systolic-mac-array systolic-mac-array Public

    Parameterizable INT8 CNN accelerator (systolic MAC array, Conv1->Conv2->FC) -- self-checking testbench + Python golden model verified over 1,900 cases, CI on every push.

    Verilog