SignMate: an Indian Sign Language translation app born from a childhood gap
By Thakur Prashanth · October 15, 2026 · 5 min read
Siddharth Sadu (BTF-015) and Naga Nidesh (BTF-018) had never used GitHub. They trained their own ML model on Indian Sign Language and built SignMate, a two-way ISL translation app now used at a school for deaf children.
At a glance
Product: SignMate
Built by: Siddharth Sadu (BTF-015), Naga Nidesh (BTF-018)
Background: Siddharth Sadu: 3rd year engineering, MVSR Engineering College; Naga Nidesh: Engineering, Matrusri Engineering College
Cohort: Build Theory Fellowship, Cohort 01 (BT-01)
Status: In use at a school for deaf and hard-of-hearing children; more pilots planned.
The personal connection
Siddharth Sadu's mother works at a school for deaf and hard-of-hearing children. Growing up, Siddharth visited the school often but avoided talking to the students, not out of embarrassment but because he did not know sign language. Learning it felt slow, and the gap could leave the children feeling left out.
The product
With Naga Nidesh (BTF-018), Siddharth (BTF-015) built SignMate: an Indian Sign Language (ISL) translation app for two-way communication between someone who signs and someone who does not. It translates ISL into text and voice, and text or voice back for the signer. It is designed for deaf students in Classes 6 to 10.
From no GitHub to training an ML model
Neither had built a product or project before, or used a GitHub repository. During the fellowship they trained and fine-tuned their own machine-learning model on ISL signs.
Where it is now
SignMate is ready and already used by people visiting the school where Siddharth's mother works. The team wants to pilot it with more deaf and hard-of-hearing people and more schools. They describe it as aiming to be India's first ISL translation app.
What this story shows
Personal experience can reveal a problem that matters, but the product has to serve the people living it. The team went from zero repository experience to training a domain-specific ML model.
Frequently asked questions
What is SignMate?
SignMate is an Indian Sign Language translation app that converts ISL to text and voice, and text or voice back, built by Siddharth Sadu (BTF-015) and Naga Nidesh (BTF-018).
Where is SignMate used?
At a school for deaf and hard-of-hearing children, with more pilots planned.
Did they train their own model?
Yes. They trained and fine-tuned a machine-learning model on Indian Sign Language signs.
Build your own product in Cohort 02
This story is part of Build Theory Cohort 01: what 24 fellows built. Build Theory is a 30-day online fellowship, based in Hyderabad, where students and early professionals go from a real problem to a live product with 1:1 mentorship and AI tools. Any year, any branch. Check your readiness score or apply for Cohort 02.
- #signmate
- #indian sign language app
- #isl translation
- #assistive technology for deaf students
- #train ml model
- #siddharth sadu
- #naga nidesh
- #btf-015
- #btf-018
- #build theory fellow stories
- #build theory cohort 01
- #student founders india
- #ai product building
- #build theory fellowship