Computer Engineering @ NUST CEME · Pakistan
Founder & lead developer of ProGenEDA · Founder & lead developer at Type2Learn
Building editable native EDA projects and accessible, typing-based active learning.
Name : Muhammad Taha Bin Zaeem
Location : Pakistan
University : NUST CEME
Field : Computer Engineering
Build Mode : AI + circuit simulation + architecture + research pipelines
North Star : Turn hard technical ideas into testable engineering artifactsI am not trying to look like a generic developer profile.
I build from the uncomfortable edge: where AI has to create real files, where simulators refuse to cooperate, where assembly becomes a full project, where research needs logs instead of vibes, and where a prototype must survive actual testing.
I like systems that are hard enough to expose weak thinking.
|
EDA automation that turns supported circuit intent into editable native project files. Natural language goes in. Strict circuit IR is produced. Local validation and repair run. A downloadable Explore: Website · GitHub organization Why it matters: this is not another chatbot wrapper. It is an attempt to make AI produce simulator-ready engineering artifacts.
|
Accessible active learning built around typing, learner choice, and clear feedback. Type2Learn pairs structured courses with learner-controlled supports, including narration, alternative input, and focus controls. Typing is a way to show learning—not a speed test. Explore: Website · GitHub organization Why it matters: accessible learning should preserve learner agency while making active participation practical.
|
|
A functional chess engine written in MIPS assembly. Board representation, coordinate move input, move validation, special rules, check/checkmate detection, and a simple AI opponent. Why it matters: assembly stops being theory when it has to hold a game state and enforce rules.
|
A local-first context and workflow system. A serious app architecture around project files, context packs, searchable local data, explicit user approval, redaction, provider routing, and reproducible handoff flows. Why it matters: AI workflows need memory, structure, consent, and traceability — not just another text box.
|
|
Research pipeline built around reproducibility. Raw sources, deterministic scripts, prompt templates, stored LLM outputs, metadata, checksums, and environment records. Why it matters: if a paper depends on LLM outputs, the pipeline must preserve the evidence trail.
|
A local workflow for turning opaque binaries into readable technical maps. It preserves folder structure, runs controlled analysis, writes structured outputs, records failures, and treats tooling as a repeatable engineering process instead of a one-time stunt. Why it matters: real computer engineering means being able to move from black-box behavior toward understandable structure.
|
ProGenEDA
Type2Learn| Signal | Evidence |
|---|---|
| 🔥 I build beyond coursework | AI circuit generation, MIPS chess, context systems, research pipelines |
| ⚙️ I care about real artifacts | .pdsprj output, Docker deployment, SQLite storage, generated manifests |
| 🧪 I care about reproducibility | checksums, stored outputs, metadata, locked baselines, audit docs |
| 🧱 I like low-level thinking | MIPS assembly, simulator design, memory layout, architecture experiments |
| 🌐 I can ship products too | ProGenEDA native-file workflows and Type2Learn accessible active-learning experiences |
| Repository | What it represents |
|---|---|
| ProGenEDA | EDA automation platform for editable native circuit project files |
| Type2Learn | Accessible, typing-based active learning |
autodecom |
Local automation workflow for structured technical analysis outputs |
Mips_Chess_Engine |
Chess engine built in MIPS assembly |
PROJECTINFINITY |
Local-first AI context/workflow system |
CS-117-Project |
Single-cycle Verilog CPU with scalar and vector operations |
FOP-Project |
C++ symbolic algebra solver with an AST-based simplification pipeline |
| Short Term | Make AI-generated simulation files reliable enough to trust. |
| Medium Term | Build publishable research pipelines with clean evidence trails. |
| Long Term | Become the kind of engineer who can move between software, hardware, AI, and systems without fear. |
rules:
- build first, polish second, document before forgetting
- vague ideas are uncompiled specifications
- if a system cannot be tested, it is still mostly imagination
- if research cannot be audited, it is only a story
- do not worship tools; make them serve the goal