York University CS · entering year 3, Fall 2026

ML Researcher.
Hardware Engineer.

Classifying gastric ultrasound images to predict anesthesia risk before surgery. Building sub-quadratic arithmetic on FPGA for AI workloads. Two tracks, one focus: precision where it matters.

Selected work kiankhooban8800@gmail.com →
Active research
Gastric Ultrasound CNN
Classifying stomach content using Perlas 3-point scale (Grade 0-3) from antrum cross-sectional area measurements.
PyTorch + MONAI + U-Net
Hardware research
B4G3 on Artix-7 FPGA
Implementing B4G3 floating-point multiplication with explicit CARRY4 instantiation to resolve delay penalties in AI inference.
SystemVerilog + Vivado
Stack
Python PyTorch MONAI SystemVerilog React SQL C++

Selected Work

Projects across medical AI, hardware design, and civic technology.

Civic Tech · Hackathon
Arrive

AI-powered settlement navigator for refugees and immigrants in Canada. Built for the IBM Z x UNSA Sheridan Hackathon 2026. Connects newcomers to legal aid, settlement services, and support organizations through a watsonx.ai-powered search interface.

IBM Z Hackathon 2026 React + Vite watsonx.ai Vercel
GitHub →
ML Research
Gastric Ultrasound CNN

Segmenting the gastric antrum in ultrasound images and classifying stomach content on the Perlas scale to assess aspiration risk before anesthesia in emergency surgery.

Undergraduate Research PyTorch MONAI U-Net Kaggle GPU
In progress · Prof. N. Mohaghegh
Digital Design · SystemVerilog
Booth Encoder Radix-2

Complete 8x8 signed Booth multiplier in SystemVerilog. Covers partial product generation, mux logic, and full summation with signed arithmetic. Three testbenches for exhaustive coverage.

SystemVerilog Testbench Signed Arithmetic FPGA-ready
GitHub →
Desktop App · macOS
Sticky-AI

macOS sticky notes powered by AI. Built in Electron with Gemini 1.5 Flash, the app extracts tasks from freeform text and organizes them automatically. Ships a Liquid Glass-inspired UI design.

Electron Gemini 1.5 Flash JavaScript macOS
GitHub →
Also in the lab
Verilog ALU (4-bit, 8 ops, overflow detection), 4x4 partial products multiplier, WiFi scanner in C++.
All repos on GitHub →

Two Research Tracks

Medical imaging and AI hardware. Supervised undergraduate research at York University on both fronts.

Track 01: Medical AI

Perioperative Gastric Imaging

Training a CNN pipeline to classify stomach content from ultrasound images of the gastric antrum. The classification directly informs anesthesia risk decisions before emergency surgery, using the Perlas 3-point grading system.

Grade 0: empty stomach
Grade 1: fluid in right lateral decubitus only
Grade 2: fluid in both positions
Grade 3: solid content
Volume: 27.0 + (14.6 × CSA_RLD) − (1.28 × Age)

Supervised by Prof. Navid Mohaghegh, York University Lassonde School of Engineering.
Dataset sourced from Dr. Anahi Perlas group, Toronto Western Hospital.

Track 02: AI Hardware

B4G3 Arithmetic on FPGA

Implementing B4G3 (Block 4-to-3 Grundy) floating-point multiplication on the Xilinx Artix-7 FPGA using explicit CARRY4 chain instantiation. The target: resolve the ASIC delay penalties identified in Langhammer et al. (ARITH 2024) and test zero-skipping under sparse AI workloads.

Baseline: B4G3 on Artix-7, explicit CARRY4 HDL
Extension: zero-encoding skipping for sparse inputs
Target venues: FPT, ARITH short paper
Tools: SystemVerilog, Vivado, Artix-7

Supervised by Prof. AmirAli Amirsoleimani, York University.
Based on Langhammer et al. ARITH 2024 reference implementation.

Tehran to Toronto.
Building carefully.

Arrived in Canada in 2021 from Tehran. Now completing my third year of CS at York University's Lassonde School of Engineering, running two undergraduate research projects simultaneously.

I work late and think slowly. Most of what I build comes from trying to understand something precisely enough to make it work in code. The gastric ultrasound project started with a professor's failing pipeline. The hardware project started with a paper and a chip.

Targeting the Vector Institute for research involvement and the Anthropic Fellowship for the 2027 cohort. Attending Shopify Builder Sundays in Toronto when the registration window opens.

School York University, Lassonde School of Engineering
Program BSc Computer Science, Year 3 (Fall 2026)
Languages English, Farsi (Persian)
Location Toronto, Ontario, Canada
Certs Azure DP-900, AZ-900 (in progress) · CS50 SQL
Targets Vector Institute · Anthropic Fellowship 2027 · DS/ML internship
Open to Research roles, ML engineering, hardware-adjacent positions in Toronto
Current tools
Python PyTorch MONAI SystemVerilog Vivado React JavaScript TypeScript SQL C++ Verilog HDL Next.js FastAPI Docker Git Azure Kaggle CVAT Electron MS SQL Server

Let's work together.

Open to research collaborations, ML engineering roles, and conversations about medical imaging, AI hardware, or anything at the intersection of technology and consequential problems.

kiankhooban8800@gmail.com