● Developed an LLM-based AI agent for automated documentation of the bank’s microservices, reducing time to write documentation from 2 days to 3 hours● Developed a script for writing credit documentation, expected to cut document processing costs by 10x (in progress)

● Analysed the impact of Fetcherr's price recommendation algorithm on key customer KPIs● Automated ETL processes to create data marts for the Analytics team● Created analytics dashboards for the revenue management team

● Responsible for developing ML models for customer segmentation based on CRM data (used k-means clustering and decision trees). Increased marketing email CTR by 300%.● Gathered functional requirements from our client and converted them into product hypotheses● Completed training in recommender systems, working with SQL databases and delivered apresentation on Interpretable ML

● Created a prototype of a mental health assistant/ smart diary● Conducted 50+ customer development interviews● Tested product hypotheses (personal profile with recommendations, rule-based chatbot, smartdiary)

● Built a regression model for predicting depression from blood tests. Used PCA and gradient● boosting to improve results.● Integrated and analysed data from multiple biosensors, developed software for a scientificexperiment

As part of UCL Robotics, I have organised a Neurotechnology conference covering such themes as Brain-Computer interfaces, noninvasive brain stimulation, new types of electrodes for recording brain activity and analysing neural population data.I have proposed and advertised the conference, introduced the subject to attendees and facilitated discussion. The conference was attended by researchers from Oxford University, UCL, Imperial College London and the Francis Crick Institute.

Completed a 10-week research project funded by the Laidlaw Scholarship as well as leadership development training. As part of the project, I have worked closely with Prof. Beth Jefferies and her team to design a behavioural experiment investigating semantic cognition (our ability to remember and use conceptual knowledge). This has resulted in a succesful pilot study. I have also analysed neuroimaging data using specialised software (FSL) and Python scripts. I have conducted a principal component analysis to explore the data and presented it to Prof. Jefferies who used it in a grant application. Working in the lab and completing the leadership development program has been very useful for improving my communication skills as part of a team.

This was a simple part-time job where I entered data from written surveys into Excel for a scientific laboratory. The most useful outcome of this experience was connecting with members of the lab with whom I continued to work as part of the Laidlaw scholarship.