PhD Researcher in AI & Computer Vision

KU Leuven (MeBioS Division)

Researching AI-based agricultural insect monitoring using computer vision and deep learning.

Upwork (currently not taking new clients)

As a freelancer specialising in computer vision and machine learning, I help clients turn their computer vision projects into life. View my profile.

• • I am building realtime deployable computer vision based surveillance systems that are deployable.
• • Technologies Using: Vast.ai, Python, PyTorch, OpenCV, Ultralytics, Django

Chair of Computer Science 7 - FAU Erlangen-Nuremberg

As a student assistant specialising in cybersecurity, I am building exercise materials for the course Introduction to Cybersecurity Fundamentals for Networking.

• • I built exercise to mimic real-world cyber attacks, helping students understand attack vectors and defense mechanisms.
• • Technologies Using: Kali, Metasploit, Wireshark, Scapy, Burp Suite, TCPDump, VirtualBox

Primetals Technologies GmbH

As a working student specialising in machine learning, I focus on finding algorithms to detect anomaly events to monitor the metal production process using concepts like Instance Segmentation, Object Detection, Keypoint Matching, and Anomaly Detection.

• • I built a web tool to evaluate a model’s performance by visualizing an event’s occurrence, which allowed us to perform the model’s evaluation on 100s of hours of videos in just a few minutes.
• • I developed a novel algorithm to detect the vibration of a static camera based on the camera frames, allowing our models to predict events with high confidence.
• • I built several synthetic data generator tools mimicking the real environment to train OCR, segmentation models, and detection models, which allowed us to generate rare event samples for training.
• • Technologies Using: Python, OpenCV, PyTorch, ONNX, TensorRT, FastAPI, Streamlit, MySQL, Docker, and Git

Upwork

As a top-rated freelancer, I helped clients turn their data applications into life.

• • I built a dynamic alert-based strategy back-testing platform and trading bots for options and stocks using Polygon.ai, Alpaca, and Interactive Brokers API.
• • Technologies Used: Python, FastAPI, Streamlit, MySQL, Asyncio, Websockets, and Git

Extenso Data Pvt. Ltd.

As an Associate Data Scientist/Engineer working in a Fintech, I was responsible for building a data pipeline to collect, process, and analyze financial data from various sources.

• • I built a backbone for an intelligent model to automate data pipelining, analysis, and reporting of fintech data for business stakeholders to make timely customer decisions
• • I built an Intelligence System to visualize the performance of loans and prospects for stakeholders to make timely decisions.
• • Technologies Used: Python, MySQL, Pentaho, Apache Airflow, Apache Superset, and Git

Ensemble-Matrix

As a Junior ML Engineer, I did experiments to efficiently detect forged signatures and built an Intelligence System.

• • Using Siamese-type neural networks, I developed an algorithm to create a synthetic hand signature to train a model to detect forged signatures.
• • Technologies Used: Python, Google Colab, OpenCV, Keras, Docker, and Git

diyo.ai

As a Unity 3D Intern, I learned how to build AR applications.

• • I designed and built a user flow for an application to improve clients’ jewelry business using Google AR Core and Unity 3D.
• • Technologies Used: Unity3D, and Git

MPercept Technology

As an AI Developer Intern, I learned how to build and deploy AI models, focusing on computer vision and chatbots applications.

• • Developed and maintained conversational agents for platforms like e-commerce, travel agency, and academic consultancy.
• • Technologies Used: Python, Rasa, OpenCV, Google Colab, AWS EC2, and Git

KU Leuven, Belgium (MeBioS Division) · KU Leuven Profile

• Research on AI-based agricultural insect monitoring.
• Promoter: Prof. Wouter Saeys

Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany

• Advanced studies in machine learning, deep learning, computer vision, and statistical analysis. Focus on practical applications of AI in real-world scenarios.
• Thesis: Enhancing Smart Grid Security: A Deep Learning Approach to Adversarial Intrusion Detection · Code · PDF
• Supervisors: Dr.-Ing. Loui Al Sardy, Mamdouh Muhammad, M.Sc.

Tribhuvan University, Nepal

• Foundation in computer science, mathematics, and programming and early exposure to machine learning and software engineering concepts.
• Academic Project: Devanagari Word/Character Recognition