komal Sukheja

I'm a Machine Learning Engineer focused on computer vision and deep learning — I care about building models that actually run in production, not just pass evaluation metrics.

My core strengths are object detection, image classification, and deploying interactive ML applications. Here's what I've shipped recently:

Face Mask Detection System (YOLOv8, OpenCV, Python)

I custom-trained a YOLOv8 architecture to classify Mask vs. No Mask in real time, with colour-coded bounding boxes for instant visual auditing. The system handles both webcam and video file inputs, uses frame-skipping logic to maintain smooth inference under load, and supports optional output video saving for historical review. The focus was on performance and scalability — not just accuracy in a notebook.

Animal Image Classifier (EfficientNetB0, PyTorch, Gradio, Hugging Face)

Fine-tuned EfficientNetB0 (pre-trained on ImageNet) for multi-class animal species classification, including a filtering layer that rejects non-animal inputs cleanly. Built an interactive Gradio interface with Top-3 match predictions and real-time confidence scores for model transparency. Deployed and hosted entirely on Hugging Face Spaces — optimised to run on standard CPUs without speed compromise.

Beyond these, I've also built a Smart Document Scanner (OpenCV, perspective transformation, Streamlit), a Customer Churn Prediction app (TensorFlow ANN, 85% accuracy), and a predictive Olympics Data Analysis platform — all deployed as interactive web applications.

My stack: Python · PyTorch · TensorFlow · YOLOv8 · OpenCV · EfficientNet · Keras · Hugging Face · Gradio · scikit-learn · Roboflow · Docker · Streamlit · Flask · Pandas · NumPy · SQL · Power BI

I studied Accounting & Finance at IBA Sukkur — which I think makes me a stronger engineer. I understand what decisions downstream teams need to make with model output, and I build with that in mind. I'm currently completing a Diploma in Computer Science.

I'm actively looking for a fully remote junior ML or computer vision role. I move fast, I learn faster, and I build things that work outside the lab.