Computer Engineering Student at UET Taxila. Engineering at the boundary of Generative AI systems, high-performance C++ software, and foundational digital hardware logic.
I am a Computer Engineering student (3rd semester) at UET Taxila with a rigorous grounding in C++, driven by low-level system design, algorithmic problem-solving, and machine intelligence.
With solid grounding in computer systems and modern engineering principles, I enjoy deconstructing challenging problems into clean, robust code. My technical journey bridges the entire computing continuum—from discrete semiconductor diode logic and transistor circuitry up to cutting-edge Generative AI pipelines, local LLM orchestration, and parameter-efficient model adaptations.
Rigorous engineering curriculum combining hardware theory, mathematical modeling, and modern software paradigms.
CURRENT: 3RD SEMESTER
Intensive real-world immersion in state-of-the-art Generative AI frameworks, vector databases, and parameter-efficient fine-tuning.
Completed an intensive 8-week corporate training and internship program focused on end-to-end generative AI pipeline development, transformer model adaptations, and semantic retrieval systems.
Architected semantic retrieval systems using FAISS index structures, contextual chunking strategies, and sentence-transformers.
Trained Low-Rank Adaptation (LoRA) layers on Llama architectures to adapt foundation models efficiently for specialized domains.
Deployed local inference runtimes with Ollama, building interactive Streamlit workspaces for real-time response streaming.
Coupled Whisper automatic speech recognition (ASR) with lightweight reasoning models for low-latency voice AI workflows.
A curated comic portfolio spanning intelligent neural pipelines, local inference tooling, and discrete semiconductor circuit design.
Production-ready Retrieval-Augmented Generation system designed for clinical query resolution. Ingests biomedical corpora into FAISS vector space with TinyLlama context synthesis.
Privacy-first local LLM chat workstation featuring multi-model hot-switching, token streaming, context memory buffers, and custom system prompt presets.
Parameter-Efficient Fine-Tuning (PEFT) framework targeting Llama-3-8B. Injects low-rank decomposition matrices into transformer attention heads for medical dialog reasoning.
Asynchronous pipeline piping OpenAI Whisper acoustic speech transcription directly into TinyLlama logic nodes for immediate structured task reasoning and command execution.
Pixel-perfect streaming portal clone featuring dynamic catalog carousels, responsive grid transitions, modal preview dialogues, and clean semantic architecture.
Console-based fitness logging application built in C++ for a Programming Fundamentals Open-Ended Lab. Uses parallel arrays to track exercises, computes workout volume and estimated calories, identifies the best lift via linear search, and persists sessions via file I/O.
Constructed fundamental boolean logic gates via discrete semiconductor diodes and pull-up/down resistors. Verified voltage thresholds and truth table fidelity on breadboard.
Designed a two-stage touch-activated switch using cascaded BC547 NPN transistors in a Darlington-style configuration, detecting the human body's leakage current (1–10 µA) to drive an LED or buzzer output. Verified switching behavior and current calculations on a breadboard prototype.
Studied and analyzed the ADC0804 8-bit analog-to-digital converter as part of Digital Logic Design coursework — mapping input voltage to digital output, the role of Chip Select and Read control signals, and how reference voltage and clock RC components affect conversion timing and resolution.
Built a browser-based webcam app that detects facial expressions and hand/pose gestures in real-time using MediaPipe, triggering matching meme images and captions. Includes a configurable gesture-to-meme mapping system for easily adding new triggers without touching the core detection logic.
Bridging software engineering, machine learning pipelines, and discrete electronic hardware design.
Whether you have a compelling project, research collaboration, internship opportunity, or simply want to discuss GenAI pipelines and systems engineering — my inbox is always open.