Skip to content
Awais Asghar.

●Portfolio · 2026 · AI Engineer, NUST

AI engineer working on ML, LLMs and AI agents.

Machine learning, deep learning, LLM apps and agents, from research models to systems that run on real hardware.

View work Download resume
projects shipped
41
projects shipped
disciplines
8
disciplines
CGPA at NUST
3.63
CGPA at NUST
Portrait of Awais Asghar

Awais Asghar

GitHub

Now: open to AI / ML roles and research Available

01Selected work

Eight AI projects, from research to deployment.

Deep learning research, computer vision, LLM systems, and models taken all the way to edge hardware.

All 41 projects
LLMs2026

AI Exam Paper Generator (RAG)

Turns lecture slides into structured exam papers with FAISS retrieval and Llama 3.1.

  • Python
  • LangChain
  • FAISS
  • Groq API
AI / ML2025

RecycleVision

EfficientNetB0 transfer learning sorts waste into six recyclable categories.

  • Python
  • TensorFlow
  • Keras
  • EfficientNetB0

02About

From a notebook to real hardware.

I graduated in 2026 with a BE in Electrical Engineering from NUST, Islamabad, with a 3.63 CGPA and a 3.94 GPA in my specialization. My final year project was an FPGA hardware accelerator for real-time image segmentation: a custom U-Net compressed 56 times, quantized to INT8, and run on a Zybo Z7-20 at under 3 W.

Most of my work lives where machine learning meets constrained hardware. I have trained segmentation and classification models, deployed classical vision on a Jetson Nano, written RISC-V processors in SystemVerilog, and built RTOS firmware on STM32. Lately I have been shipping LLM products too: a RAG exam generator, a chat backend on Vercel, and a receipt-reading ledger app.

In summer 2026 I was a Mitacs Globalink research intern at the NRC Herzberg Astronomy and Astrophysics Research Centre in Victoria, Canada, training a spatiotemporal network to find faint Trans-Neptunian Objects in telescope image sequences.

03Experience

Research labs, a chip design centre, and a classroom.

  1. Summer 2026

    Victoria, BC, Canada

    Mitacs Globalink Research Intern

    NRC Herzberg Astronomy & Astrophysics Research Centre · with Dr. Sébastien Fabbro

    • Applied ML research on large astrophysical image datasets: built a variable-frame spatiotemporal U-Net that replaces classical shift-and-stack detection of Trans-Neptunian Objects.
    • Ran the full research loop on the CANFAR science platform: synthetic object injection with real PSFs, HDF5 data pipelines, training, evaluation and documentation.
    PyTorchAstronomyCANFARHDF5

  2. Feb 2025 – May 2026

    Islamabad, Pakistan

    Chip Design Trainee Engineer

    NUST Chip Design Centre (NCDC) · with Dr. Hammad M. Cheema

    • RTL design, simulation and FPGA implementation in SystemVerilog; hands-on with RISC-V, computer architecture, Linux and C.
    • Implemented single-cycle and 5-stage pipelined RV32I processors with hazard detection and forwarding, verified on Artix-7 at 100 MHz.
    • Designed a reprogrammable FPGA anti-theft car security system with debounced sensors, siren generation and a fuel-pump interlock.
    SystemVerilogRISC-VVivadoQuartus

  3. June 2024 – Sept 2024

    Islamabad, Pakistan

    Deep Learning Research Intern

    Deep Learning Lab, SINES, NUST · with Dr. Jameel Nawaz Malik

    • Built and evaluated CNN and transfer-learning models in PyTorch and TensorFlow with CUDA-accelerated training, focused on benchmarking and validation.
    PyTorchTensorFlowCUDA

  4. May 2024 – Aug 2024

    Islamabad, Pakistan

    Embedded Systems Design Research Intern

    ESDAC Lab, NUST · with Dr. Usman Zabit

    • Developed and deployed Embedded Linux on FPGA and DE1-SoC boards: bootloaders, kernel configuration, GPIO, memory management and real-time processing.
    Embedded LinuxDE1-SoCVerilogQuartus

  5. Aug 2024

    Islamabad, Pakistan

    Robotics & AI Tutor

    Murabbi, NUST · with Dr. Wajahat Hussain

    • Co-led a robotics and AI boot camp using EV3 kits and model training. Students built competition-winning sumo, speed and obstacle-avoiding robots and later presented them to the Prime Minister of Pakistan.
    TeachingRoboticsMentoring

  6. June 2023 – Aug 2023

    Islamabad, Pakistan

    Machine Learning Research Intern

    HamsanTech, NSTP · with Dr. Nazia Perwaiz

    • Built a multi-model skin cancer classification pipeline (XGBoost, AdaBoost, LightGBM, SVM, logistic regression) reaching 91% accuracy and 0.963 AUC from a single RGB image.
    • Delivered Power BI and Tableau dashboards and exploratory analysis on real industry datasets.
    Scikit-learnXGBoostPower BITableau

04Toolkit

What I reach for.

Machine learning & AI

  • Python
  • PyTorch
  • TensorFlow
  • Keras
  • Scikit-learn
  • NumPy
  • Pandas
  • OpenCV
  • SciPy
  • CUDA

LLMs & RAG

  • LangChain
  • Groq API
  • Llama 3
  • FAISS
  • CrewAI
  • Prompt engineering
  • Vector search

FPGA & digital design

  • SystemVerilog
  • Verilog
  • RISC-V
  • Vivado
  • Vitis HLS
  • Quartus
  • ModelSim
  • Zybo Z7-20
  • Artix-7
  • DE1-SoC

Embedded systems

  • STM32
  • ESP32
  • FreeRTOS
  • Arduino / AVR
  • Jetson Nano
  • Pixhawk
  • Embedded C
  • UART / I2C / SPI

Languages

  • Python
  • C / C++
  • Embedded C
  • TypeScript
  • MATLAB / Simulink
  • Assembly
  • CUDA

Data & tools

  • Power BI
  • Tableau
  • Excel
  • Matplotlib
  • Seaborn
  • Git / GitHub
  • Linux
  • Jupyter
  • LaTeX
  • Vercel

Certifications & coursework

  • Machine Learning Specialization (3 courses)DeepLearning.AI
  • Deep Learning Specialization (5 courses)DeepLearning.AI
  • Introduction to Deep Learning, 6.S191MIT
  • Advanced Python ProgrammingUdemy
  • Machine Learning, Deep Learning, Computer Vision, Computer Architecture, Digital System Design, DSP, Embedded System Design, Control Systems, RISC-V, LinuxNUST & NCDC coursework

Leadership

  • Millennium Fellow, BreatheSafe Initiative

    UN Academic Impact & Millennium Campus Network

    Leading a community project that deploys low-cost air quality monitors for dust, smoke and harmful gases, to protect vulnerable populations and push for cleaner air.

  • Director Outreach

    Google Developer Solution Club, NUST

    Ran outreach and partnerships for student developer events.

  • Founder, CashFlow

    Startup · 2026

    Building a personal finance app that reads receipts and bank statements with AI and keeps loans and savings out of income. Live at usecashflow.vercel.app.

05Recognition

Papers in progress and honors.

Publications

  1. Design of a Novel Lightweight U-Net Architecture for Efficient Semantic Segmentation

    In progress

    Based on the FPGA accelerator FYP and CARLA driving-scene segmentation work at NUST SEECS.

  2. Real-Time Fabric Defect Detection on Edge Devices Using Classical Computer Vision

    In progress

    Multi-method GLCM / FFT / Gabor pipeline with IoU fusion, deployed on Jetson Nano.

Honors & awards

  • 2026

    Mitacs Globalink Research Internship

    Fully funded international research placement at NRC Herzberg, Canada, summer 2026.

  • 2025

    Millennium Fellowship (Class of 2025)

    Selected among the top 4% globally by the UN Academic Impact and Millennium Campus Network.

  • 2024

    NUST SEECS Merit Scholarship

    Awarded for academic performance, 2024–2026.

  • 2022–2026

    Dean's Honor List

    Recognized four times for outstanding academic performance.

  • 2025

    Prime Minister Youth Laptop Scheme

    Merit-based award for academic excellence.

  • 2022

    3rd place, BISE Multan Pre-Engineering

    Ranked third among 70,000+ students.

  • 2020

    Intermediate Scholarship

    100% fee waiver for academic excellence.

06Contact

Let's build something.

Roles, research collaborations, or a question about one of the projects. Messages land straight in my inbox.