About Me
Highly motivated and results-oriented Senior Data Scientist with a proven track record of delivering innovative solutions across diverse industries. Particular expertise in generative AI, large language models (LLMs), na…
Highly motivated and results-oriented Senior Data Scientist with a proven track record of delivering innovative solutions across diverse industries. Particular expertise in generative AI, large language models (LLMs), natural language processing (NLP), MLOps.
Experience
Senior Data Scientist
<p>Independently built a generative (and non-generative) AI chatbot framework which can be used to build, test and deploy production-ready chatbots with little-to-no code needed.<br>
This library will be used for all generative AI chatbots within the bank and acts as a proprietary alternative to Langchain.<br>
Designed a state-of-the-art model hosting platform within Kubernetes for hosting LLMs in a sustainable, scalable and highly optimised manner.<br>
Delivered presentations to key stakeholders within the bank to raise awareness of how generative AI can be leveraged to drive efficiency and reduce risk across the bank.<br>
End-to-end design, training and deployment of a proprietary large action model.<br>
This model was used as an orchestrator for a multi-agent virtual assistant that is capable of fulfilling complex requests.</p>
Senior Data Scientist
● Independently built a generative (and non-generative) AI chatbot framework which can be used to build, test and deploy production-ready chatbots with little-to-no code needed. This library will be used for all generative AI chatbots within the bank and acts as a proprietary alternative to Langchain.
● Designed a state-of-the-art model hosting platform within Kubernetes for hosting LLMs in a sustainable, scalable and highly optimised manner.
● Delivered presentations to key stakeholders within the bank to raise awareness of how generative AI can be leveraged to drive efficiency and reduce risk across the bank.
● End-to-end design, training and deployment of a proprietary large action model. This model was used as an orchestrator for a multi-agent virtual assistant that is capable of fulfilling complex requests.
Data Scientist
<p>Designed and implemented RegGPT, a GPT model that has been fine tuned on regulatory text (SFT + RLHF) for document information retrieval.<br>
Improved OCR accuracy for complex scanned regulatory documents.<br>
Created an ensemble language detection model which outperformed all existing open-source models.<br>
Reduced inference time across the entire ML pipeline by 70% through the use of ONNX, Sparsification and the deployment of Triton Inference Server to Kubernetes.</p>
Lead Data Scientist
<p>Built an automated report system that sends personalised weekly reports to senior fraud managers at more than 30 telcos (e.g. Vodafone, EE, BT, Orange and Virgin).<br>
Developed an NLP-based data mining algorithm that searches the internet and identifies reports of Wangiri fraud.<br>
Created an anomaly detection system using ensemble learning to predict the evolving strategies of Wangiri fraud.<br>
Designed and implemented a highly available and scalable microservice architecture in GKE.</p>