Zain Ali

• Computer vision and deep learning
• Medical image analysis
• Multimodal learning (vision, audio, and language)
• Model reliability, calibration, and interpretability
• Evaluation methodology for health and social-impact applications

BS Computer Science·CGPA 3.65 / 4.00

Taxila, Pakistan·Graduated July 2026

Coursework: Computer Vision, Artificial Intelligence, Artificial Neural Networks, Big Data Analytics, Linear Algebra, Probability & Statistics.

Final Year Project: PrepArt: A real-time multimodal mock-interview system that fuses video, audio, and text signals to generate structured interview feedback. Advised by Dr. Qamas Gul.

• Section LeaderatStanford Code in Place (CS106A)

  2025 & 2026

  • Selected as a section leader for Stanford’s global introductory Python course, taught from the CS106A curriculum.
  • Led weekly live sections for 12+ international students, teaching control flow, decomposition, data structures, and debugging.
  • Gave individual feedback on assignments and held office hours, focusing on getting students unstuck rather than handing over answers.

• Prime Minister’s Youth Laptop Scheme

  Awarded on academic merit at UET Taxila.

  2025

• Harvard CS50x Puzzle Day: Winner

  Perfect 9/9 solve in both editions.

  2025 & 2026

• UC Berkeley CALICO

  Placed 126th out of 476 teams.

  2025

• Meta Hacker Cup: Round 1 Qualifier

  Advanced past the qualification round.

  2024

• Advent of Code

  Completed all 50/50 stars.

  2024

• LeetCode

  200+ problems solved across data structures and algorithms.

  Ongoing

• Prism.ai

  2026

  A six-stage AI research pipeline that takes a raw dataset from analysis through literature retrieval, experiment planning, and code generation to a written research report.

  PyTorch · FastAPI · Next.js · OpenAlex · GPT-5.6

• PrepArt

  2026

  Final year project: a real-time multimodal mock-interview system that fuses facial and eye cues, vocal prosody, and transcript analysis into structured interview feedback.

  Python · MediaPipe · Librosa · MoviePy · PyTorch · FastAPI

• Brain Tumor MRI Classification

  2025

  Four-class brain tumor classification from MRI using MobileNetV2 transfer learning, with Grad-CAM used to check that the model attends to tumor regions rather than scan artifacts.

  PyTorch · MobileNetV2 · Grad-CAM · scikit-learn · NumPy

• Skin Lesion Classification

  2025

  Seven-class dermatoscopic lesion classification on HAM10000 (~10k images) with ResNet-18 transfer learning and class-weighted loss to counter severe class imbalance.

  PyTorch · ResNet-18 · HAM10000 · torchvision · scikit-learn

• Full Stack DeveloperatGhetch OS

  Jun 2025 - Present

  Remote

  • Build and ship end-to-end product features across a TypeScript/React front end and a Python service layer.
  • Design REST APIs and data models, and own the integration work between the client, backend services, and third-party providers.
  • Improve page performance and reliability through query optimisation, caching, and instrumentation of hot paths.
  • Work asynchronously with a distributed team, reviewing code and maintaining deployment workflows.

• Backend DeveloperatFireprenair

  Aug 2024 - Apr 2025

  Remote

  • Developed and maintained backend services and REST APIs powering the core product.
  • Modelled relational schemas and wrote migrations, keeping data access patterns efficient as the dataset grew.
  • Integrated authentication, payments, and external APIs, with error handling and logging around each boundary.
  • Wrote tests and participated in code review to keep releases predictable.

Languages

Python, TypeScript, JavaScript, C++, SQL

Machine Learning

PyTorch, torchvision, scikit-learn, NumPy, Pandas, Transfer learning, Grad-CAM

Vision & Multimodal

OpenCV, MediaPipe, Librosa, MoviePy, Image classification

Web & Backend

FastAPI, Next.js, React, Node.js, REST APIs, PostgreSQL

Tools

Git, Docker, Linux, Jupyter