Skip to content
View mhussainahmad's full-sized avatar
:octocat:
Learning
:octocat:
Learning

Highlights

  • Pro

Block or report mhussainahmad

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

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

Report abuse
mhussainahmad/README.md

Muhammad Hussain Ahmad

ML and robotics engineer, four years across robot control, perception and model inference. M.Sc. student in Mechanical Engineering (Robotics) at the University of Manitoba, Winnipeg; B.S. Mechanical Engineering, NUST.

My research is markerless teleoperation of a 7-DoF manipulator (Kinova Gen3): RealSense and MediaPipe RGB-D perception at 100 Hz driving the arm in real time, with occlusion and dropout handling, gesture intent detection, a Genesis digital twin with payload estimation, and a 1 kHz C++ torque controller and safety layer underneath. That code is not public yet.

I also built a causal audio-visual emotion recognition model for real-time streaming: 75.9% accuracy at 6 ms per step, distilled from an offline model (paper accepted at GLOBCER).

Public repos

Project What it is
gauntlet Regression testing and failure analysis for learned robot policies. Runs a policy across seeded grids of perturbations (lighting, camera pose, clutter, latency, sensor corruption, dust, glare) on MuJoCo, PyBullet, Genesis or a numpy crop-row env, and reports where it fails, with confidence intervals and paired comparisons. On PyPI as gauntlet-robotics.
flux FP8-quantized FLUX.1-schnell inference pipeline, tuned for fast generation on an RTX 4090.
Chicken-disease-classification-project End-to-end image classifier: DVC pipeline, VGG16 transfer learning, Flask serving, Docker and CI/CD.
face_detection / kivy-app Siamese-network face verification, and a Kivy desktop app that runs it on a webcam.
Classical ML projects
Project What it is
Online-Payments-Fraud-Detection Decision-tree classifier flagging fraudulent mobile-money transactions.
Stock-price-LSTM Stock price forecasting with an LSTM.
Future_Sales_Prediction Sales prediction from ad spend across platforms.
Waiter_Tips_Prediction Linear regression predicting restaurant tips, with Plotly exploration.
Pfizer-Vaccine-Sentiment-Analysis VADER sentiment scoring and word clouds on Pfizer vaccine tweets.
winequality-prediction ElasticNet on red wine quality, with a DVC pipeline, Flask serving and CI.

Experience

  • Research Assistant, Robot Control and Safety, University of Manitoba (2024–present)
  • Robotics Engineer, Teleworker AI (2026): ROS 2 scanning pipeline on a Unitree Go2, running on the onboard NVIDIA Jetson Orin
  • ML Engineer, Wombo (2024): SDXL and FLUX inference optimization
  • Freelance ML Engineer (2022–2024)

Contact

Pinned Loading

  1. gauntlet gauntlet Public

    Robustness evaluation harness for robot manipulation policies (VLA): perturbation sweeps, failure clustering, reproducible reports.

    Python

  2. flux flux Public

    Python

  3. Chicken-disease-classification-project Chicken-disease-classification-project Public

    Jupyter Notebook

  4. face_detection face_detection Public

    This is a deep learning algorithm focused on detecting the face using siamese neural network.

    Jupyter Notebook