AI Research Engineer · MSc Artificial Intelligence

RAYSON
FERNANDES

Building intelligent systems at the intersection of Deep Learning, Molecular Biology, and Agentic AI — currently at Brunel University London.

View Projects Get in Touch
SCROLL

Who I Am

Building the future of AI

I'm an AI Research Engineer and MSc Artificial Intelligence candidate at Brunel University London, specializing in Deep Learning, NLP, and Generative AI systems.

Currently contributing to the BARMIE research project — an open-source computational pipeline for ecological hazard ranking of endocrine disrupting chemicals — in collaboration with the University of Southampton.

I architect end-to-end ML pipelines, build autonomous agentic workflows with LangChain and crewAI, and deploy production-ready AI systems on AWS and Azure.

When I'm not training models, I'm designing RAG solutions, fine-tuning transformers, and exploring the frontiers of AI safety and interpretability.

3+
Years ML experience
95%
Detection accuracy (voice AI)
40+
Regressors benchmarked
881
PubChem features engineered

Technical Arsenal

Core Skills

🧠 Deep Learning
PyTorch TensorFlow Transformers Multi-Head Attention Rotary Embeddings ResNet CNN RNN
🤖 Generative AI & Agents
LangChain crewAI RAG Vector Databases Knowledge Graphs LLM Fine-tuning
⚗️ ML & Data Science
Scikit-learn XGBoost LightGBM Random Forest Pandas NLP Feature Engineering
☁️ Cloud & MLOps
AWS Azure Docker Kubernetes CI/CD Microservices Streamlit
💻 Languages & Tools
Python SQL Git Power BI RDKit Librosa ChEMBL API
🔬 Bioinformatics
Protein-Ligand Docking Binding Affinity EC50 / IC50 Lipinski Descriptors PaDEL 3D Biomolecular

Career

Work Experience

AI Research Assistant
Jan 2026 – Present
Brunel University London
  • Contributing to the BARMIE research paper in collaboration with the University of Southampton — an open-source pipeline for ecological hazard ranking of endocrine disrupting chemicals.
  • Designed end-to-end deep learning pipelines to predict drug toxicity and molecular potency by mapping large-scale 3D protein-ligand datasets to binding affinity metrics (EC50, IC50).
  • Conducting rigorous research on model architectures and hyperparameter optimization to improve predictive performance and generalization.
Jr. Data Scientist Intern
Feb 2025 – May 2025
Codanto Goa
  • Implemented predictive models and analytical workflows on real-world data science projects.
  • Managed end-to-end data lifecycle: collection, cleaning, preprocessing, and feature engineering.
  • Designed interactive visualisations and reports to communicate insights to stakeholders.
  • Applied CI/CD pipelines, version control, and reproducible workflow best practices.
Machine Learning Intern
Oct 2023 – Nov 2023
CoE Digital Forensics
  • Achieved 95% detection accuracy in AI voice manipulation by engineering a bespoke PyTorch + Librosa preprocessing pipeline.
  • Reduced audio data noise by 30% and optimised feature extraction latency significantly.
  • Analysed audio patterns to distinguish real from AI-generated speech with high precision.

Portfolio

Featured Projects

01 / Research
High Potency Molecule Prediction for Influenza A (PB2)
Extracted bioactivity data via ChEMBL API, engineered 881 PubChem Fingerprint features with RDKit/PaDEL, and benchmarked 40+ regressors using LazyPredict to predict compound potency.
Python RDKit XGBoost LightGBM Scikit-learn ChEMBL API
02 / Deep Learning
AI Voice Manipulation Detection System
Hybrid deep learning model combining AudioMiniEncoder with a Transformer classification head. Implemented Multi-Head Attention, ResNet blocks, Rotary Embeddings, and ALiBi Bias with a Streamlit real-time interface.
PyTorch Librosa Transformers Streamlit Torchaudio
03 / Computer Vision
Image Classification with VGG-Inspired CNNs
Custom VGG16/VGG19-inspired architecture in PyTorch and TensorFlow for Fashion MNIST multi-class classification. Achieved 90% accuracy with Conv2D, MaxPooling, Dropout, and Adam optimizer.
PyTorch TensorFlow CNN Fashion MNIST
04 / Machine Learning
House Price Prediction with Regression Models
End-to-end regression pipeline for real estate price prediction. Fine-tuned hyperparameters and handled data imbalances to deliver high-accuracy analytical insights for property decision-making.
Python Scikit-learn Pandas Regression
View All on GitHub →

Academic Background

Education

MSc in Artificial Intelligence
Brunel University of London
January 2026 – Present
Deep Learning Machine Learning Neural Networks Predictive Data Analysis
B.E. in AI and Data Science
Sahyadri College of Engineering & Management
November 2021 – July 2025
Data Science & Big Data Business Intelligence Data Visualisation Database Management

Let's Connect

Get in Touch

I'm open to research collaborations, AI engineering roles, and interesting ML projects. Whether you have a research question, job opportunity, or just want to talk about AI — I'd love to hear from you.

Send an Email
✉
Email Raysonfernandes955@gmail.com
in
LinkedIn rayson-fernandes-056440225
⌥
GitHub github.com/Rayson-fernandes
◎
Phone +44 7345 683567