Experience · CV

Where I've worked

01 · General information
Full nameHemanth Sai GarladinneLocationHyderabad, Telangana, IndiaLanguagesEnglish, Hindi, German (limited working)Top skillsEdge deployment, Research & Development (R&D), Large Language Models (LLM)
02 · Experience
Apr 2026 – Present
✦
Voice AI Systems Engineer — Qualcomm
Voice AI · Hyderabad, Telangana
Jan 2024 – Apr 2026
Havells India Ltd
2 years 4 months · Bengaluru, Karnataka
Associate Research EngineerJul 2025 – Apr 2026
  • Designed, trained and deployed Indian-accented ASR systems for ESP32-S3 microcontrollers: real-time inference in a 6 MB footprint. Built a novel 1.4M-parameter model at 17% WER / 9% CER on 1,000 hours of Indian English.
  • Implemented and benchmarked Conformer, Conmer and stateless RNN-T architectures. Built an automated audio data pipeline with GPT-4o and Azure/Sarvam TTS, growing domain data from 6 to 54 hours. Contributed custom TensorFlow Lite Micro operators.
  • Built a RAG platform over enterprise documents for AI-powered search and Q&A across the organization.
  • Two-stage retrieval (dense search, cross-encoder reranking, metadata filtering): 82% Recall@k, 95% Answer Relevance, 88% Answer Faithfulness on an internal benchmark.
  • End-to-end LLM observability with Phoenix: latency, retrieval quality and generation performance in production.
Graduate Engineering TraineeJul 2024 – Jul 2025
IoT InternJan 2024 – Jul 2024
Jul 2022 – Oct 2022
NLP Intern — IVIS Labs Private Limited
Internship completion certified

Built a question-generation application in PyTorch. Context understanding via sense2vec and KeyBERT, F1 score 0.84.

03 · Education
2020 – 2024
✦
B.Tech, Artificial Intelligence & Machine Learning
Vellore Institute of Technology, Bhopal
04 · Honors & awards
Top 10, Meta Llama Hackathon ↗Winner, Docker AI/ML Hackathon ↗PyTorch Docathon H1 Winner
05 · Certifications & publications
06 · Open source projects
Add your open-source contributions here. (Old site listed the al-folio theme as a template entry.)
07 · Academic & other interests
Speech recognition, NLP, low-latency voice interfaces, multilingual speech, edge ML. Hobbies: reading research papers, trail running.