Your responsibilities:

- Build and ship production-ready AI/ML featuresfrom data ingestion and feature engineering to model training, evaluation, and deployment.
- Develop LLM/GenAI solutions (prompt engineering, tool use, guardrails) and RAG pipelines (chunking, embeddings, vector search, caching, re-ranking).
- Optimise training and inference performance via batching, quantisation, distillation, LoRA/PEFT, accelerator utilisation (GPU/TPU), and efficient memory/latency tuning.
- Build and maintain MLOps/LLMOps workflows—CI/CD for models and prompts, model registry/versioning, feature stores, and automated promotion across environments.
- Instrument observability for data, models, and prompts (telemetry, metrics, traces, dashboards, alerts); implement A/B tests and online/offline evaluation.
-
Embed Responsible AI considerations (fairness, explainability, safety, bias testing) and document assumptions, datasets, and limitations.

- Document architecture, workflows, and best practices to support scalability and ongoing maintainability.
- Conduct code reviews, write unit/integration/e2e tests (including data and prompt tests), and uphold engineering standards and documentation.
- Work with advanced AI/ML frameworks, cloud services, and container orchestration platforms.
- As an AI Engineer, you are responsible for designing, building, and deploying scalable AI and machine learning solutions that solve realworld business problems, partnering closely with data scientists to productionize models and integrate them seamlessly into applications and enterprise workflows

Your Profile

Essential skills/knowledge/experience:

AI Engineer (5 to 12 Years)

- Hands-on experience with GenAI, Gemini or Open source LLMs , Train , finetune and Onboard new LLMs




- Experience in building GenAI applications using Python
- Hands-on Experience with API Development and Microservices architecture and End to End integrations
- Knowledge of RAG (Retrieval-Augmented Generation ) and ADK, MCP
- Solid understanding of LLMs, prompt engineering, and graph-based workflows.
- Hands-on Experience with API Development and Microservices architecture
- Experience in CI/CD pipelines, and containerization (Docker/Kubernetes)., Harness and Git actions.
- Practical experience implementing LLM and GenAI solutions, including prompt engineering, model fine-tuning, RAG pipelines, embeddings, and vector databases.
- Build scalable data pipelines and workflows on GCP (Big Query, Vertex AI, Dataflow, Pub/Sub, Redis and NoSQL Databases , Maintaining chat history etc.
- Optimize model performance, monitor production systems, and ensure reliability , Auto Scaling using Prometheus, Dynatrace and Lang Smith

Desirable skills/knowledge/experience: (As applicable)

- Solid hands-on experience building and deploying machine learning models, including preprocessing, feature engineering, training, evaluation, and optimisation.
- Knowledge of API Gateways and ISTIO , ability to Diagnose and intercept failures in End to End communication.
- Implement best practices for data governance, security, and MLOps on GCP.
- Proficiency with Python and common AI/ML frameworks such as TensorFlow, PyTorch, JAX, scikit-learn, and Hugging Face libraries.




- Knowledge of MLOps and LLMOps practices—including CI/CD for models, model registry/versioning, feature stores, orchestration, and automated deployments.
- Ensure AI solutions meet security, privacy, compliance, and responsible AI standards.
- Understanding of secure engineering and data protection practices, including IAM, secrets management, encryption, and safe handling of sensitive data.
- Ability to optimise performance of training and inference pipelines—profiling, quantisation, distillation, batching, caching, or hardware acceleration.
- Collaborate with data scientists to productionize models and integrate them into applications, workflows, and APIs.

Questions

- End-to-end delivery: "Walk me through an AI/ML feature you took from data ingestion and feature engineering all the way to deployment.
- LLM/GenAI & RAG depth: "Describe a RAG pipeline you built. How did you handle chunking, embeddings, vector search, re-ranking, and caching — and how did you measure answer quality?"
- Performance optimization: "Tell me about a time you cut inference latency or training cost. Which techniques did you use — quantization, distillation, LoRA/PEFT, batching, GPU/TPU utilization — and what was the measurable gain?"
- MLOps & observability: "How have you set up CI/CD for models and prompts, model versioning/registry, and observability (metrics, traces, alerts)? How did you run A/B tests or online/offline evaluation to catch regressions?"
- Responsible AI & collaboration: "Give an example where you addressed fairness, bias, explainability, or safety in a model. How did you partner with data scientists to productionize their work and document assumptions and limitations?"

📌 AI Engineer (Hyderabad)
🏢 Tata Consultancy Services
📍 Hyderabad