Welcome
Currently, I’m a PhD student at Harvard, supervised by Hima Lakkaraju, researching trustworthy ML. My current interests include efficient data attribution techniques and faithful reasoning in LLMs.
I previously worked within JPMorgan AI Research under the direction of Daniele Magazzeni and Saumitra Mishra, where I researched new global counterfactual explanation methods and their impact in assessing model fairness.
I’m a graduate of the Cambridge MEng, where I was supervised by Adrian Weller and Umang Bhatt. My research centered around providing meaningful explanations for uncertainty estimates from probabilistic models.
Outside of work I’ve been a regular footballer, played for the University of Cambridge 1st team, coached the Corpus Christi College team, and currently compete for MIT FC in the BSSL. I once woke up and decided I wanted to run a marathon without any training. I’d like to do so again to test my mental limits.
Download my CV for further details.
- Data Attribution
- LLM Reasoning
- Interpretability / Explainability
MEng in Explainable AI, 2021
University of Cambridge
BA in Computer Engineering, 2020
University of Cambridge
Experience
- Explainable AI research on global counterfactual explanations, implementing a state-of-the-art (NeurIPS) method and identifying inefficiencies, proposing a modified method that executes 8 times faster when achieving the same level of performance
- Continuation of MEng research on explaining uncertainty in deep learning
- Training models in PyTorch for generation of counterfactual explanations
- Exploring the notion of a distribution over counterfactual explanations for a single input
- Finalising AAAI'22 submission
- Object-oriented programming in a finance setting using Python (testing with pytest), Flask, sklearn, tensorflow and SQL
- Planned a solution for a disaster relief charity to port 40% of in-person training to online training and initiated contact with a software-service company to discuss technical and financial details of our solution (£200k+ annual savings proposed)
- Co-inventor on 3 separate pending patent applications for arithmetic hardware designs with improved PPA (Power, Performance, Area) over industry standards; worked with the datapath team in an R&D environment
- Learnt to rapidly interpret code from past/current team members and make changes (Linux, Python, Perforce, VHDL)
Education
1st Year: Class I (87%, 12th of 324)
2nd Year: Class I (83%, 12th of 310)
3rd Year: Pass (COVID/No Classing)
4th Year: Honours Pass with Distinction
Masters Project in Explainable AI
Outstanding Project Award for top 5% of students in Information Engineering
1st paper accepted to ICLR workshops (first-author, travel award)
2nd paper accepted to ICML workshops (first-author)
Combined paper submitted to NeurIPS conference (under review)
Specialisation of Computer and Information Engineering
Dewhurst Scholarship for First and Second Year Results
- 4A*s: Mathematics, Furthermathematics, Physics and Chemistry
- College Award for Academic Excellence in Mathematics
- Senior Team Mathematics Challenge Regional Winners and National Final Competitors
- British Mathematical Olympiad Qualification through Senior Mathematical Challenge
- 13 A*s: Double Maths, Double English, Triple Science, Statistics, Astronomy, Spanish, French, History, Geography
- School Award for Highest Academic Achievement
Languages
Passive Fluency
Full Fluency
Conversational
Contact
- d.w.ley "at" hotmail "dot" com
- +44 7522 105139
- Corpus Christi College, Trumpington St, Cambridge, CB2 1RH
- DM Me
- Tweet Me
- Skype Me