AugmntX

AugmntX is the world’s largest network of pre-vetted talent.

Contact Info

305, IRIS Corporate Centre,
Hiranandani Meadows, 400610.
hello@augmntx.com
+919820045154

Follow Us

Home / Developers / AXO31932
profile image
Anuradha A AXO31932 Active
AI ML Developer in Mumbai
AI/ML engineer with over six years of experience in voice AI, speech processing, and LLM-driven applications. Specializes in building real-time ASR-to-LLM-to-TTS pipelines and deploying scalable AI systems on cloud platforms. Proficient in Python, Node.js, PyTorch, TensorFlow, and Hugging Face Transformers. Skilled in speech models like Whisper, wav2vec2, Tacotron, and VITS, with hands-on expertise in MLOps tools including MLflow, Weights & Biases, Docker, and Kubernetes. Experienced in audio processing, GPU optimization, and real-time communication systems, delivering multilingual voice assistants and domain-specific LLMs across agriculture, healthcare, and finance
Industries : Food Tech, AgroTech
Share
Hire Anuradha A
Download PDF
Availability
Full-time
Total experience
6 years
Technical skills
AI
7 Years
NLP
7 Years
Python
7 Years
LLM
7 Years
Vector Database
6 Years & 5 Months
Open AI
6 Years & 5 Months
Hugging face
6 Years
RAG
6 Years
Pinecone
6 Years
Playwright
6 Years
Node.js, Whisper, Tacotron, VITS, AWS Lambda, FFmpeg, MLflow, CloudWatch, PyTorch, wav2vec2, AWS S3, Lambda, Step Functions
Projects
CloudFood
March-2024 to November-2024
Description
Description Cloud Food is an AI-driven food ordering system that enables users to place and track orders using natural voice conversations. It integrates real-time ASR LLM TTS pipelines to process speech input, interpret intent, and deliver natural voice responses. The platform ensures seamless, low latency interactions through serverless deployment and MLOps driven model lifecycle management on AWS.
Roles and Responsibilities
Designed and implemented a real-time ASR LLM TTS voice pipeline for end-to-end voice-based ordering. Integrated OpenAI Whisper for accurate speech-to-text transcription and Tacotron VITS for speech synthesis. Developed contextual reasoning modules enabling the LLM to interpret multi-turn conversations. Automated MLOps workflows using MLflow for model tracking and version control. Deployed scalable, serverless components on AWS Lambda, achieving 400ms latency per request. Optimized pipeline for GPU based inference and implemented monitoring via CloudWatch. Collaborated with cross functional teams to ensure smooth CI CD integration and delivery. Conducted user-level testing to evaluate conversational quality and response time.
Technologies: Python, Node.js, Whisper, Tacotron, VITS, AWS Lambda, FFmpeg, MLflow, CloudWatch
Industry: Food Tech
AgroConnect Voice Assistant
May-2023 to November-2025
Description
AgroConnect is a multilingual AI voice assistant for farmers providing crop guidance, weather updates, and market price insights. It supports Hindi and English conversations through custom-trained ASR and domain-tuned LLMs. The system ensures reliable performance in rural network conditions via efficient AWS orchestration.
Roles and Responsibilities
Designed ASR models using wav2vec2 for Hindi and English speech recognition. Developed domain-specific LLM prompts for agricultural advisory. Integrated ASR correction layer, improving accuracy by 35%. Built AWS Step Functions workflow for ETL and model inference orchestration. Implemented model evaluation using WER and BLEU metrics. Collaborated with product teams for language dataset annotation. Deployed pipeline with scalable, event-driven architecture on AWS.
Technologies: PyTorch, Hugging Face, wav2vec2, AWS S3, Lambda, Step Functions
Industry: AgroTech
Education
Bachelors degree in Computer Science
SRTM University
2015 to 2019
Language
English - Intermediate