ML Platform & Fullstack AI Engineer

ML Platform & Fullstack AI Engineer

Years of experience - 4+

Location - Hyderabad

Hybrid Model

This role builds and maintains the foundation to allow productionizing AI use cases easily. Because there is no existing AI platform or standardized way to integrate AI products into our internal production systems, this is not a maintenance role; it is a platform-building role. The T-shape means depth in ML infrastructure and AI platform engineering, with the horizontal reach to build the integration layer between AI systems and existing production systems.

CORE RESPONSIBILITIES

– Build and maintain the ML platform from the ground up: experiment tracking, model registry, model serving, CI/CD for ML, and monitoring

– Own GenAI infrastructure: vector databases, LLM serving (e.g. LiteLLM), RAG pipeline architecture, evaluation and observability frameworks

– Design and implement data pipelines that feed both classical ML models and GenAI applications into production

– Own the integration layer

— wire AI outputs reliably into existing systems

– Support defining infrastructure standards and best practices that existing engineers and future hires can build against

KEY SKILLS

– ML infrastructure: MLflow / W&B, model serving frameworks, feature pipelines, CI/CD for ML

– GenAI infrastructure: vector stores (Pinecone, Weaviate, pgvector), LLM serving, evaluation frameworks (RAGAS, LangSmith)

– Cloud ML services: AWS SageMaker preferred – Strong software engineering fundamentals: API design, distributed systems, containerisation (Docker/Kubernetes)

– Systems integration: REST APIs, message queues, connecting ML services to enterprise systems

WHAT GOOD LOOKS LIKE

– Has built an ML platform or substantial part of one in a production environment — not just used one

– Has productionized both a classical ML models and a GenAI application (e.g. RAG system or LLM-powered feature) end-to-end

– Thinks in systems considers reliability, scalability, and observability before writing the first line of code

– Strong enough as a software engineer to own integration work without needing a dedicated backend team