Bemnet Beshah
Software engineer and machine learning researcher.
I study computer science and mathematics at Texas State University and research data augmentation for sign-language recognition.
- Studying
- Computer Science and Mathematics, with a minor in Physics, Texas State University
- From
- Addis Ababa, Ethiopia
- Based in
- Austin, Texas
- Research
- Machine learning: computer vision and training data for sign recognition
Research
See research notesCan body-model augmentations turn a small set of sign videos into a larger, useful training set for sign recognition?
Labeled sign-language video is scarce, so I am building a way to generate extra training clips that keep each sign's meaning.
Read the project- 20-clip round trip: 3.13% body error, 5.10 mm hand mean
- Face geometry added; first augmentation trials found arm changes 3–4× too strong
- nextTrain the same recognizer with and without the augmented clips
Projects
See all projectsInterview Prep MCP Server
A hosted MCP server that gives Claude and ChatGPT one shared, per-user spaced-repetition study database.
KareApp Invoice Builder
Turns a food distributor's messy delivery spreadsheet into a Zoho import, with Claude name matching and human review of every match.
HSP Vendor Finder
A browser agent searches Texas's HUB vendor list by trade code, checks each vendor, and saves up to three state info sheets per trade.
Achievements
See all achievements- 2025[MLH] Best Use of Gemini 2.5 Computer Use Model, HackTX 2025Ayn, built with a team of four
- AI4ALL IgnitePresented the wind energy site suitability project at the program symposium
Writing
See all postsWhat a 20-clip round trip says about sign video geometry
A field note on a 20-clip WLASL round-trip assessment: what body and hand reconstruction measurements support, and what they leave unanswered.