AI / ML Engineer • Computer Vision • GenAI • Cloud
Master’s student in Advanced Computer Science at Newcastle University. Focused on GPU-optimized inference, scalable ML pipelines, and Agentic workflows.
AI/ML + Cloud-first engineering, focused on performance and deployability.
Graduate Software Engineering / Data Science roles (UK/Global). I like building complete solutions: data → model → app → deployment.
Academic background & professional certifications.
Specialization in Machine Learning, Deep Learning, Distributed Algorithms, and IoT. Distinction.
First Division (7.11/10). Focus on Neural Networks, CV, and Big Data Analytics.
Self-directed study using open academic resources (MIT, Harvard, freeCodeCamp).
Core tools I use to build, train, and deploy.
From raw data to scalable production deployments.
Cleaning, Augmentation, Feature Engineering
Architecture Selection, Transfer Learning
Quantization, Pruning, Latency Reduction
Containerization, API Serving, Monitoring
I optimize training pipelines for maximum throughput using Mixed Precision and custom CUDA kernels.
Deploying efficient architectures (EfficientNetV2) with interpretability layers.
View EvidenceBuilding context-aware agents using LangChain and Vector Search to reduce hallucinations.
View EvidenceServing models via Dockerized FastAPI endpoints with high-concurrency optimizations.
View EvidenceSelected work and open source contributions.
Diagnosing retinal diseases from OCT scans with high accuracy and interpretability.
Ensemble of EfficientNetV2 for feature extraction and XGBoost for classification, augmented with Grad-CAM.
High-precision diagnosis with visual explainability maps for clinicians.
Early detection of crop diseases to prevent yield loss in agricultural settings.
End-to-end CNN pipeline optimized for edge deployment, integrated with Azure cloud services.
Scalable, real-time inference API ready for mobile integration.
Generating context-aware, engaging captions for social media content automatically.
Retrieval-Augmented Generation (RAG) using LangChain and Transformer models for context understanding.
Contextually relevant captions with continuous evaluation loop.
Bridging the gap between research models and production-grade serving infrastructure.
Containerized microservices (Docker) serving ONNX models via FastAPI with CUDA acceleration.
High-throughput, low-latency inference endpoints.
Get in touch via email or phone.
