Mark Edmonds Email & Phone Number

Boston, Ma, Us

San Francisco, California, Us

- Imitation and reinforcement learning for Trajectory Generation.- Distillation for on-board model inference- Human preference learning

San Francisco, California, Us

Los Angeles, California, Us

Humans build generalizable and explainable representations of their environment through interaction, observation, imitation, intervention, and language. The following research produced from a collaborative, team­-based lab explores how artificial agents can use these five concepts to learn robust and transferable representations of tasks and environments. During my PhD, I have worked on 4 major areas:1) Imitation learning: Training a robot to twist open a medicine bottle2) Causal learning: Virtual escape room to examine how humans and AI learn transferable causal representations3) Explainable AI: How can robots explain their behavior to foster human trust? 4) Representation learning: The role of language in building generalizable representationsFor details on these projects, please see my Resume or CV (linked in Featured section of my LinkedIn profile).

Santa Monica, Ca, Us

- Taught one­/two 45­ student classes/quarter: lead lectures, hold office hours, and create course materials.- Instructed 25 courses: Internet Programming (HTML, CSS, JavaScript, MySQL, and PHP), Intro to C, and Intro to C++.

- Worked on representation learning project to couple vision and language.- Built a virtual environment using Unreal Engine 4 (UE4) for embodied AI that couples language (scene graph) and vision.- Using the environment to create a dataset consisting of images, scene graphs (language labels), and object segmentation.- The goal of this environment is to provide a playground to learn generalizable, commonsense representations where the latent representation is tightly coupled with language labels.

- Created and hosted an hour-­long weekly interview and live performance podcast with 29K+ listens.- Interviewed 32 electronic music artists including multiple Grammy­ nominated guests.- Edited and released over 70 episodes, DJ'­ed 40 live mixes, created branding and marketing material to build fanbase.

Los Angeles, Ca, Us

- Teaching assistant for Introduction to C, Introduction to C++: lead discussion hours (∼50 students), held office hours

Wright-Patterson Afb, Oh, Us

- Conducted Declarative Memory research as part of the ACT-R cognitive architecture.- Used CUDA to parallelize declarative retrievals, yielding a 100x speedup overthe fastest existing implementation.- Implemented thread pools, parsers, IPC.

Olathe, Ks, Us

- Interned as a member of the Datalink team in the Aviation Department.- Contributed to the ACARS protocol for Garmin Avionics software.- Reduced verification testing time by 40%.

Doctor Of Philosophy (Ph.D.), Computer Science

Master Of Science (M.S.), Computer Science

Bachelor Of Science (B.S.), Computer Engineering