Ai Ml Engineer

Adecco India; Chennai District; On-site; INR 1,800,000 - 2,000,000; Full time

Job summary

Adecco India is seeking a Senior AI Engineer in Chennai to design, build, and evaluate LLM-based systems that automate healthcare administrative workflows. You will own prompt architecture, model selection, RAG pipelines, and verification layers to meet production-grade accuracy and cost targets.

The role focuses on reducing latency and cost per transaction while ensuring robust evaluation and integration with data engineering for training data and go-live readiness.

Qualifications

• 5+ years of experience in ML/AI engineering.
• Proficient with TensorFlow or PyTorch.
• Experience with MLOps and production deployment.
• Experience with Generative AI, LLM fine-tuning, and RAG.

Responsibilities

• Design and own prompt/agent architecture and task decomposition for production healthcare-automation workflows.
• Select, benchmark, and tune large language models against defined targets for accuracy and cost.
• Build and maintain a retrieval-augmented generation (RAG) pipeline: embeddings, vector-store retrieval, grounding outputs.
• Develop an evaluation harness with golden datasets and automated scoring.
• Design an independent verification/guardrail layer for model outputs.
• Own and drive down cost per transaction and latency; balance hosted vs self-hosted models.
• Collaborate with data engineering on training/test set construction and with stakeholders on go-live criteria.

Skills

Advanced ML/DL Modeling

MLOps & Production Deployment

Generative AI, LLM Fine-tuning, RAG

Solution Architecture & ML Problem Sol

Job description

Urgent Hiring !!!!

Senior AI Engineer

• Openings: 1
• Experience: 5+ Years
• Location: Chennai
• Salary Range 18 to 20 LPA

Mandatory Skills

• Advanced ML/DL Modeling (TensorFlow/PyTorch)
• MLOps & Production Deployment
• Generative AI, LLM Fine-tuning, RAG
• Solution Architecture & ML Problem Solving

About the Role

Design, build, and evaluate LLM-based and agentic systems that automate complex healthcare administrative workflows - turning a well-defined business process into a production-grade system with a measured accuracy and cost bar, not a prototype or demo.

Key Responsibilities

• Design and own prompt/agent architecture and task decomposition for a production healthcare-automation workflow, breaking a multi-step business process into steps a model can reliably execute and be evaluated against.
• Select, benchmark, and tune large language models against defined accuracy, precision/recall, and cost-per-transaction targets.
• Build and maintain a retrieval-augmented generation (RAG) pipeline: embeddings, chunking strategy (fixed, hierarchical, or semantic), vector-store retrieval, and grounding of model outputs in retrieved evidence.
• Build and maintain an evaluation harness - golden datasets, automated scoring, precision/recall/F1 tracking - that proves accuracy before any production release.
• Design and implement an independent verification/guardrail layer that checks model outputs before they are acted on.
• Own and drive down cost per transaction and latency in production; make model-selection tradeoffs between hosted APIs and self-hosted models.
• Partner with data engineering on training/test set construction and with quality stakeholders on go-live criteria.