Overview

About Reliance Intelligence

Responsibilities

• Build and deploy on-device ML models for JioFrames AI features — object detection, scene understanding, gesture recognition, OCR, and health sensing
• Optimize models for wearable SoC deployment using quantization (INT8/INT4), pruning, and knowledge distillation within strict power and memory budgets
• Develop and maintain real-time inference pipelines handling camera frames, microphone input, and sensor fusion on constrained hardware
• Build and fine-tune NLP models for intent recognition, wake-word detection, and command parsing running locally on JioFrames AI
• Integrate LLM capabilities for the JioAI assistant — enabling natural conversation, live translation, and contextual Q&A
• Establish ML training pipelines, experiment tracking, and model versioning using MLflow or equivalent
• Design and implement OTA model update mechanisms to continuously improve AI capabilities post launch
• Collaborate closely with firmware, hardware, and Android teams to ship AI features end-to-end

Qualifications

• 4–10 years of hands-on ML/AI engineering experience with production deployments
• Strong proficiency in PyTorch or TensorFlow and on-device inference runtimes — TFLite, ONNX Runtime, or QNN SDK
• Proven experience in model optimization for edge deployment — quantization, pruning, and latency profiling
• Solid background in computer vision — object detection, classification, OCR, or depth estimation
• Experience with NLP or multimodal AI; familiarity with LLM fine-tuning or RAG pipelines is a plus
• Comfortable working close to hardware — NPU/DSP pipelines, camera ISP, and wearable power constraints
• Strong Python skills; C/C++ and experience with Qualcomm Hexagon SDK is advantageous
• Experience shipping AI features on mobile, wearables, IoT, or embedded platforms preferred