Umesh Timalsina

Lead System Engineer,
Vanderbilt University

I’m a full-stack research software engineer with 7+ years in learning technology and multimodal analytics. I’ve led systems for U.S. classroom studies and an OECD trial across hundreds of European schools, and I turn ideas from learning scientists and teachers into working research software, fast.

Timeline of projects from 2019 to the present. Select a project to read more.
  1. DeepForge DeepForge, 2019 – 2022
  2. GMSO GMSO & MoSDeF, 2019 – 2022
  3. Symbench Symbench design tools, 2019 – 2022
  4. ChimeraPy ChimeraPy, 2022 – 2023
  5. Hull design Surrogate-based hull design optimization, 2022 – 2024
  6. SyncFlow SyncFlow, 2023 – present
  7. Betty's Math Betty's Math, 2025 – present
  8. Classroom AI Real-time classroom AI feedback, 2025 – present
  • Learning technology
  • Research software
  • Engineering design and ML

About

I’m a full-stack research software engineer with 7+ years of experience building learning technologies and real-time multimodal analytics platforms. I’m a lead system engineer at ICONS (formerly the Institute for Software Integrated Systems) at Vanderbilt University, where I work with the Open-Ended Learning Environments lab and the NSF AI Institute for Engaged Learning.

I created and lead Betty’s Math, an assessment and AI-tutoring platform used in an OECD randomized controlled trial across hundreds of schools in four European countries. I also lead SyncFlow, an open-source platform that has captured 400+ hours of multimodal classroom data in U.S. studies. I like taking ideas from learning scientists, teachers and international partners and turning them into dependable research software.

More about me

Selected projects

2025 – present

Betty's Math

Creator and lead engineer

A math assessment and AI-tutoring platform embedded in the OECD's PILA and deployed in a randomized controlled trial across hundreds of schools in France, Germany, the Slovak Republic and Latvia. It includes an LLM tutor with learner-model context, LLM translation and knowledge-graph pipelines, xAPI logging and a real-time teacher dashboard.

2023 – present

SyncFlow

Creator and lead engineer

An open-source WebRTC platform for real-time multimodal classroom data. It sustains 200 concurrent streams at about 200 ms latency and has archived 400+ hours (more than 10 TB) for 1,000+ researchers and students. It is deployed in classroom studies in Indiana, Tennessee and North Carolina.

2022 – 2023

ChimeraPy

Co-developer

A distributed streaming framework for real-time multimodal data retrieval and processing, used by 100+ people in Nashville-area classroom studies. I built its orchestration layer, SvelteKit dashboard and pipeline library (video, audio, people tracking, live transcription and video anonymization).

All projects

Selected publications

SyncFlow: A Scalable Platform for Multimodal Learning Analytics

U. Timalsina, E. Davalos, N. P. Sanda, Y. Zhang, J. H. Fonteles, T. S. Ashwin, et al.

US Research Software Engineer Conference (US-RSE), 2025

ChimeraPy: A Scientific Distributed Streaming Framework for Real-time Multimodal Data Retrieval and Processing

E. Davalos, U. Timalsina, Y. Zhang, J. Wu, J. H. Fonteles, G. Biswas

IEEE International Conference on Big Data (BigData), pp. 201–206, 2023

DeepForge for Astronomy: Deep Learning SDSS Redshifts from Images

U. Timalsina, B. Broll, K. Moore, T. Budavári, Á. Lédeczi

Astronomy and Computing 40, 100601, 2022

A Machine Learning Gateway for Scientific Workflow Design

B. Broll, U. Timalsina, P. Völgyesi, T. Budavári, Á. Lédeczi

Scientific Programming 2020, 8867380, 2020

All publications

Writing

All posts

Contact

I’m glad to hear about research collaborations, classroom data infrastructure, and roles where I can build learning technology at scale.

umesh.timalsina26@gmail.com