Ahmed Aman Ibrahim

Google Scholar | Github | LinkedIn

Experienced in computer vision, NLP, and deep learning with 4+ years in applied AI and research roles. Published author with a strong academic foundation (MSc, BSc with Distinction) and a track record of delivering ML solutions for healthcare, insurance, and environmental monitoring. Skilled in designing and deploying end-to-end ML pipelines using PyTorch, TensorFlow, Scikit-Learn, and AWS. Passionate about using AI to solve real-world problems.

Experience

Research Assistant in Applied AI

University of Huddersfield | November 2024 – Present

Built and deployed regression models and graph neural networks (GNNs) using Scikit-Learn, PyTorch Geometric, Flask, and Azure to predict soil organic carbon content.
Designed and implemented Bayesian Networks and knowledge graphs (Neo4j) to model and analyze causal relationships affecting soil carbon levels. Create visualisations using networkx, matplotlib and plotly.

Machine Learning Researcher (part-time)

Center for Visual Computing, University of Bradford | July 2021 – Present

Researched and developed deep learning models for facial beauty prediction, creating novel loss functions and achieving state-of-the-art results presented at IEEE SKIMA conference in 2023.
Built and optimized person re-identification (ReID) and Single Shot Detector (SSD) models for image-to-image search in large forensic datasets.
Trained and deployed gait recognition, face verification, and face pose estimation models for forensic applications in collaboration with Dubai Police.
Train and deploy transformer embedding models for low resource languages (Sinhala and Dhivehi) to be used in RAG applications.

Artificial Intelligence Consultant (contract)

Allied Insurance Company of the Maldives | September 2024 – March 2025

Built a personalised recommendation engine that has 95%+ accuracy using customer purchase histories and demographic data using collaborative filtering. Make visualisations using Power BI.
Designed and deployed a statistical anomaly detection pipeline to identify potential insurance fraud cases which analyse customer behaviour and relationships.
Developed a custom chatbot using LlamaIndex, Claude, and OpenAI APIs, integrating internal knowledge bases, agents and internal APIs to automate customer support.
Developed and deployed an OCR automation framework using AWS Textract, reducing manual claims processing workload by 70%.

Machine Learning Engineer

Aiberry Inc. | October 2021 – October 2023

Fine-tuned LLaMA 2 and MiniLM models to analyze speech and text data for automated mental health screening, achieving 80% accuracy (state-of-the-art).
Developed multimodal ML pipelines combining vision (Facial Action Units) and NLP, enhancing depression and anxiety detection models.

Machine Learning Engineer Intern

Allied Insurance Company of the Maldives | August 2021 – January 2022

Trained and deployed convolutional neural networks (CNNs) for automated vehicle damage detection from motor accident images.
Designed a sentiment analysis dashboard for LiveChat customer data, enabling data-driven insights into customer experience trends.

Blog

Scikit-Learn Model Deployment with ONNX: Run Anywhere, Faster

Published on July 19, 2025

Stop letting pickle version conflicts break your deployments. This step-by-step guide shows you how to convert Scikit-Learn models to the ONNX format for robust, portable, and hardware-accelerated inference that runs on any platform.