[PDF] davidad_cv.pdf - davidad (David A. Dalrymple)

David A. Dalrymple

T +1 22 4444 9614

@davidad

davidad

davidad

B [davidad@alum.mit.edu](mailto:davidad@alum.mit.edu)

Formal Education

2011–2012 Harvard University, Partial Ph.D., Biophysics Took most of the neuroscience classes.
2008–2010 Massachusetts Institute of Technology, Partial Ph.D. AI under Minsky and Sussman.

2006–2008 Massachusetts Institute of Technology, S.M., Media Technology Massively parallel computing.
Youngest-ever recipient of an MIT graduate degree (age 16).
2000–2005 University of Maryland, Baltimore County, B.S., Mathematics Analysis, matrix theory, neural nets. . .

AI under Minsky and Sussman.
Massively parallel computing.

Youngest-ever recipient of an MIT graduate degree (age 16).
2000–2005 University of Maryland, Baltimore County, B.S., Mathematics Analysis, matrix theory, neural nets. . .
2000–2005 University of Maryland, Baltimore County, B.S., Computer Science Algorithms, data structures. . .

Skills and Interests

Primary areas of working knowledge: Applied category theory, AI safety, mathematical modelling, interactive
theorem proving, formal verification, modal and substructural logic, measure-theoretic probability, stochastic processes,
machine learning, nervous system imaging and simulation, information theory, decision theory, game theory, mechanism
design, numerical methods, control theory, philosophy of mind, metaethics

Computer science areas of expertise: Abstraction design (interfaces, visualizations, relational schemas, languages, or
APIs), type theory, categorical semantics, data structures (distributed, in-core, or out-of-core), databases, numerical
optimization algorithms, efficiency optimization of algorithms, data analysis, network and consensus protocols.

Selected programming skills: Rust, Rocq, C++17, C, git, ARMv8, Metal, Intel 64, (La)TeX/TikZ, PostgreSQL

Experience & Informal Education

2023–Present — Programme Director, UK Advanced Research + Invention Agency. Directing a £59M R+D
programme on Safeguarded AI, and the broader Mathematics for Safe AI ‘opportunity space’.

2021–2023 — Research Fellow, Oxford University, Faculty of Philosophy, Future of Humanity Institute. Metaethics
with an eye toward aligning AIs with moral truth (in case that exists); then shifted to a sociotechnical safety-engineering
approach which became “the Open Agency Architecture” and thence “Safeguarded AI.”

2017–2023 — Research Scientist, Protocol Labs (half-time). Co-inventor of the Filecoin protocol suite. Filecoin’s
market cap is presently over $1B, and at its peak was over $10B. More recently invented Hypercerts mechanism for
public goods funding. Contributed to research on potential future Protocol Labs projects.

2015–2017 — Intermittent software engineering & machine learning consulting; research at MIRI workshops.

2010–2011 — Independent scientist (Cambridge, MA), funded by a personal grant from Larry Page. Self-studied
probability theory and machine learning, and implemented a PCA algorithm based on 

... (truncated, 8 KB total)