John Lewis Partnership, London
Since March 2025
OctaiPipe (t-dab.ai), London E1
April 2022 - March 2025
• Contributed to 7 projects for commercial and industrial customers to deliver edge computing federated learning (FL) solutions to predict machine failures, HSE compliance and event risks, customer churns, evaluate ROI
• Main contributor in 2 Innovate UK projects: large collaborations with enterprise partners
• Experienced in real-life industrial time series, and computer vision datasets
• Implemented, benchmarked and productised state-of-the-art ML algorithms and trustworthy AI features, including federated clustering, explainability, auditability, adversarial fortification, differential privacy, leading to near-production-ready code
• Designed, implemented and deployed Retrieval Augmented Generation (RAG)/LLM infrastructure, based on LlamaIndex, FastAPI, Kubernetes, Weaviate
• Integral part of product team to deliver first version of company’s main product--- edge-IoT ML platform: Git, feature integration and codebase maintenance, continuous QA testing
• Customer-facing and communications experience--- delivered reports and presentations to stakeholders on data and model insights; disseminated knowledge with white papers and company blog posts. Customer onboarding. Produced product demo videos
• Project scoping and management experience. Working independently and collaboratively cross-team and deliver promptly in fast-paced start-up environment
Department of Physics and Astronomy
The University of Kentucky
August 2021-February 2022
New York University
Thesis Adviser: Professor Massimo Porrati
2021
The Hong Kong University of Science and Technology
First-Class Honours
2014
• Python, SQL, Git, Linux
• Machine learning and deep learning:
◦ Time series, computer vision (CV), NLP, tabular data
◦ Neural network transfer learning/fine-tuning, clustering, anomaly detection
◦ Pandas, NumPy, Scikit-Learn, SciPy, TensorFlow, Keras, PyTorch, Huggingface Transformers, LightGBM, XGBoost, statsmodels, UMAP, flower
• Trustworthy AI
◦ Explainability, differential privacy, auditability, adversarial fortification
◦ LIME, SHAP, Opacus, robust aggregation in federated learning
• Unstructured data processing: time series, image, text (NLP)
◦ Librosa for audio processing, NLTK, Gensim, Scikit-Image, OpenCV, time series statistical models for forecasting
• Infrastructure deployments and orchestration, databases--- edge IoT and cloud
◦ LlamaIndex (LLM, RAG), OpenAI, FastAPI/Flask, Streamlit, Docker, Kubernetes
◦ Azure, PySpark, InfluxDB, Weaviate vector DB
• Data visualisation, exploratory data analysis
◦ Matplotlib, Seaborn, Grafana
• Certificates: Apache Spark and Python, Microsoft Certified: Azure Data Scientist Associate, AWS SageMaker (AutoML), Neural Networks, CNN, Sequence Models, Time Series in Tensorflow
• Statistical skills, analytical skills, mathematical modelling, LaTeX, Agile
A full list of my publications in physics is available at