Skip to content
View PypCoder's full-sized avatar
:octocat:
while self.alive: self.code()
:octocat:
while self.alive: self.code()
5 followers · 9 following

Achievements

Achievement: QuickdrawAchievement: Pull Shark

Achievements

Achievement: QuickdrawAchievement: Pull Shark

Block or report PypCoder

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
PypCoder/README.md 
██████╗ ██╗   ██╗██████╗  ██████╗ ██████╗ ██████╗ ███████╗██████╗
██╔══██╗╚██╗ ██╔╝██╔══██╗██╔════╝██╔═══██╗██╔══██╗██╔════╝██╔══██╗
██████╔╝ ╚████╔╝ ██████╔╝██║     ██║   ██║██║  ██║█████╗  ██████╔╝
██╔═══╝   ╚██╔╝  ██╔═══╝ ██║     ██║   ██║██║  ██║██╔══╝  ██╔══██╗
██║        ██║   ██║     ╚██████╗╚██████╔╝██████╔╝███████╗██║  ██║
╚═╝        ╚═╝   ╚═╝      ╚═════╝ ╚═════╝ ╚═════╝ ╚══════╝╚═╝  ╚═╝

Muhammad Asad Ullah · ML/AI Engineer & Full-Stack Developer

Typing SVG

About Me

  • I'm looking to collaborate on ML/AI projects and full-stack web applications
  • I'm currently learning Rust and deepening my PyTorch knowledge (training custom architectures from scratch)
  • I'm looking for remote employment, freelance clients, and open source contributors
  • Ask me about custom neural networks, PyTorch, FastAPI, Next.js, or building full-stack AI systems
  • Fun fact: I built a neural network from scratch using only NumPy — no frameworks — and hit 97.93% accuracy on MNIST

ML / AI Stack

Frameworks & Training PyTorch NumPy Pandas Scikit-Learn Matplotlib TensorFlow Keras Hugging Face

LLM APIs & Agents Anthropic Gemini OpenAI LangChain LangGraph MCP Ollama Agentic AI RAG Prompt Engineering

Domains NLP Generative AI Anomaly Detection Fine-Tuning Bioinformatics ML Network Threat Detection

Apps & Deployment Streamlit FastAPI Google Colab

Full-Stack Stack

Backend Python FastAPI Flask REST API SQLite Supabase PostgreSQL

Frontend Next.js TypeScript React Tailwind CSS Framer Motion

DevOps & Tooling Vercel Git GitHub uv

Key Projects

Project Stack What It Does Link
MutantScope FastAPI · Next.js · AlphaFold · Gemini · NGL Full-stack protein mutation, powered by SERAPH, analysis with 3D visualization and AI explanations MutantScope
SERAPH PyTorch · CNN + BiLSTM Custom hybrid model for protein secondary structure prediction (Q3 classification) SERAPH
AI Business Analyzer Gemini · Serper.dev · SQLite · Streamlit Autonomous AI agent for competitive business analysis AI Business Analyzer
Neural Net (NumPy) Pure NumPy 97.93% MNIST accuracy — built from scratch, no frameworks Neural Network from Scratch

“As long as I’m alive, I have infinite chances.”

PypCoder Badge

Pinned Loading

  1. MutantScope MutantScope Public

    AI-powered protein mutation analysis platform. Enter any UniProt ID, mutate a residue, and SERAPH, an original ESM2+CNN+BiLSTM model, predicts exactly which secondary structures change and why.

    Jupyter Notebook 3

  2. AI-Business-Analyzer AI-Business-Analyzer Public

    AI Business Analyzer is an autonomous AI agent that turns a business idea into a structured strategic report, covering market research, competitor analysis, and growth strategy, powered by Gemini, …

    Python 5 1

  3. Neural-Network-from-Scratch Neural-Network-from-Scratch Public

    A feedforward neural network built from scratch using only NumPy. Implements forward/backprop, Adam optimizer, and multiple loss functions, achieves 97.78% accuracy on MNIST without any ML frameworks.

    Jupyter Notebook 3