InMobi - SDE III Gen AI

• Design and implement production-ready generative AI applications that serve millions of users, from initial architecture through deployment and monitoring • Build advanced RAG (Retrieval-Augmented Generation) pipelines that combine vector databases, hybrid search, and intelligent caching to deliver sub-second response times • Develop multimodal AI systems that seamlessly integrate text, vision, and audio capabilities using state-of-the-art models • Architect scalable microservices that handle thousands of concurrent AI requests while optimizing for cost, latency, and reliability • Lead code reviews and technical design sessions, establishing best practices and architectural patterns that elevate the entire team's capabilities • Optimize large language models through fine-tuning techniques to achieve domain-specific performance improvements • Implement comprehensive MLOps practices including automated testing, model versioning, A/B testing frameworks, and real-time monitoring dashboards • Collaborate with product managers and stakeholders to translate complex business requirements into innovative AI solutions • Deploy AI models across multiple cloud platforms (GCP) using containerization and orchestration technologies • Create and maintain technical documentation, runbooks, and architectural decision records that enable knowledge sharing across teams • Mentor junior engineers through pair programming, technical talks, and hands-on guidance to accelerate their growth • Research and prototype emerging AI technologies to identify opportunities for competitive advantage • Fine-tune and optimize state-of-the-art language models for specific business use cases, achieving significant improvements in accuracy and relevance • Design multi-agent AI systems using frameworks to orchestrate complex workflows and decision-making processes • Implement advanced prompt engineering strategies including Tree of Thoughts, ReAct patterns, and automatic prompt optimization to maximize model performance • Build production-grade embedding systems that handle billions of vectors, implementing efficient indexing strategies and hybrid search capabilities • Develop computer vision pipelines using models for tasks ranging from object detection to visual question answering • Create secure AI applications with robust safeguards against prompt injection, jailbreaking, and data leakage while maintaining compliance with AI governance standards • Optimize token usage and implement intelligent caching strategies to reduce costs by 50-70% while maintaining quality • Design and implement evaluation frameworks that go beyond traditional metrics, incorporating human feedback loops and domain-specific quality measures • Build real-time AI inference systems capable of processing streaming data with sub-100ms latency requirements • Integrate multiple foundation models into unified applications, implementing fallback mechanisms and load balancing for high availability • Develop custom tools and functions that extend LLM capabilities, enabling models to interact with databases, APIs, and external systems • Implement advanced RAG techniques including contextual embeddings, cross-encoder reranking, and Graph RAG for complex reasoning tasks • Create multimodal search systems that enable users to query across text, images, and documents using natural language • Build AI-powered data processing pipelines that automatically extract, transform, and enrich unstructured data at scale • Deploy edge AI solutions using frameworks like ONNX and TensorRT, optimizing models for resource-constrained environments