Eric Sheen
Building AI-powered systems
I'm a software engineer and machine learning researcher currently contributing to advancements in AI agents at Reka AI.
I'm a software engineer and machine learning researcher currently working on AI agents at Reka AI. My journey began with a passion for building systems that bridge cutting-edge technology and practical applications.
Throughout my career, I've worked on diverse technologies—from developing synthetic video generation frameworks to creating satellite image search engines. I'm particularly drawn to projects at the intersection of machine learning, full-stack development, and intuitive user experiences.
As I continue my education at the University of Waterloo, I'm focused on pushing the boundaries of what's possible with AI systems while ensuring they remain centered on human needs and values.
Life Beyond Code
In my free time, I enjoy traveling and exploring new places. Here are some moments from my journeys and personal life.
Experience
936 Labs
Building the future of video understanding with AI.
Founder
2024 - April 2025
- ›Orchestrate public market GPUs for massively distributed data processing and inference workflows significantly reducing costs.
- ›Host VLM inference API architecture and serve video generation models.
- ›Declined $100k pre-seed funding offer from Afore Capital (afore.vc) to pursue independent vision.
Related Work
Reka AI
An AI research and product company building large multimodal models.
Member of Technical Staff
Nov 2024 - May 2025
- ›Led the design and implementation of the sales agent system at Reka AI, including orchestrating backend pipelines with AWS CDK, SQS, and Lambda, and defining scalable database schemas to support dynamic agent workflows and queueing.
- ›Built full-stack infrastructure for the vertical from scratch using Next.js, Clerk for auth, and PostgreSQL, integrating real-time APIs, task management logic, and frontend tools to monitor and control agent behavior.
- ›Owned the entire lifecycle of multiple internal products, collaborating across teams to define scope, ship features fast, and ensure reliability at scale, while maintaining hands-on responsibility for infrastructure, APIs, and user experience.
DeepAI
2nd in the world for unique monthly active users for AI chat and image generation.
Machine Learning Engineer
May 2024 - August 2024
- ›Developed Simverse, a novel synthetic video generation framework for zero-shot text-to-video (T2V) generation, bypassing the need for pre-training on specific datasets using LLMs and Blender. Simverse utilizes a hierarchical JSON schema to control object placement, movement, and behavior in 3D environments. Generalized functions for advanced 3D rendering, automated camera control, and auto-captioning to maintain spatial coherence across frames.
- ›
- ›Created the Spatial-Scene-Dataset, a dataset focused on spatial positioning curated from Pexels for fine-tuning video generation models. Developed a toolkit for video preprocessing and data pipelines and secured a grant from PrimeIntellect to support this initiative.
- ›Developed the first fine-tuning scripts for CogVideo and OpenSoraPlan.
Descope
A developer-centric authentication and user management platform.
Developer Relations Engineer
May 2023 - August 2023
- ›Designed, developed, and deployed 8+ open-source sample apps using Descope authentication with React.js, Next.js, Django, Flask, and SvelteKit. Creator of the Next.js Hackathon Template soon to be featured on Vercel Templates.
- ›Utilized React.js and Descope's SDK to create Descope Explorer: an authentication search engine for users to instantly preview and interact with authentication flows, reducing development and deployment times by 500+ hours per year. One of the most actively used products across the company and by 3000+ customers.
- ›Published 9+ technical blogs, and a YouTube video, on Descope authentication, OAuth, OIDC, SSO, session management, JWTs, deep dives into front-end development concepts and rendering techniques, and sample apps.
SkyWatch
Platform providing access to satellite imagery data.
Software Developer
July 2021 - September 2022
- ›Developed a Python and TypeScript EarthCache SDK for the EarthCache API, providing SkyWatch customers with the fastest tools to integrate and access satellite data more easily in their applications.
- ›Designed and released a satellite image search engine for the Machine Learning and Sales team for labelling, reporting, and viewing images reducing the manual process time by 5.5k+ hours per year using Flask, EarthCache API, and AWS.
- ›Utilized React.js to develop and deploy a website for the EarthCache SDK documentation, blogs, and tutorials for the 200+ SkyWatch developer community.
- ›Automated the daily process of extracting and uploading data from an FTP server to an S3 bucket saving over 1000+ hours of manual upload time and human error in Python utilizing AWS.
Harvard University - Learning, Innovation, and Technology Lab
Research group focused on learning, innovation, and technology.
Student Researcher
July 2021 - July 2022
- ›Designed, developed, and deployed the MMCA project website to visualize 12+ years worth of multimodal indicators of collaboration that have been identified by social scientists over the last decade under Dr. Bertrand Schneider.
Related Work
Education
Bachelor of Software Engineering (BSE)
University of Waterloo
Software Engineering WEEF Representative, ML Research at WAT.ai
Awards & Recognition
Research Compute Grant
Prime Intellect (2024)
$200 Grant
Awarded compute resources to support independent AI research projects.
Speaker
TEDx (2019)
Sharing Ideas Worth Spreading
Invited to speak about Growth Mindset at a local TEDx event.
Grant Recipient
1517 Fund (2024)
$1000+ Grant
Received funding to support the development for video generation model finetuning.
President's Entrance Scholarship
University of Waterloo (2023)
Awarded upon admission for outstanding academic achievement.
BMO Entrance Scholarship
University of Waterloo / BMO (2023)
Awarded upon admission based on academic merit and community involvement.
Noor Monga Entrance Scholarship
University of Waterloo (2023)
Awarded to outstanding students entering the Software Engineering program.