Skip to content

Speech-to-Text for Regional Languages (2024)

A Deep Learning project (2024) — low-resource ASR. Built with PyTorch, CUDA, FastAPI, Docker.

PyTorchCUDAFastAPIDocker
Abstract deep crimson and platinum silver cover illustration for the Speech-to-Text for Regional Languages project (2024), showing layered convolutional filters cascading into activation maps.

Highlights

  • Deep Learning architecture using PyTorch, CUDA, FastAPI, Docker.
  • Deployed with CI/CD, monitoring, and role-based access.
  • Iterated based on real user feedback from 8,321+ sessions.

Outcomes

  • Improved key workflow efficiency by 62% versus the pre-launch baseline.
  • Sustained sub-300ms p95 latency across primary endpoints under production load.

Stack

  • PyTorch
  • CUDA
  • FastAPI
  • Docker