:AI Engineer(NLP)

Location : Chennai, Hyderabad and Pune

Experience : 4 to 7 Years
Design and implement NLP solutions (classification, extraction, summarization, Q&A;) aligned to product goals.
Build Retrieval-Augmented Generation (RAG) pipelines: indexing, chunking, embedding strategies, and reranking.
Develop multimodal workflows when required (text + images/structured data) and define interfaces to downstream services.
Fine-tune or adapt models (where appropriate) and establish evaluation datasets and metrics.
Must have worked with MCP based architecture.
Optimize inference for latency, throughput, and cost: batching, caching, quantization, and model selection strategies.
Implement robust prompt and tool orchestration with guardrails, monitoring, and fallbacks.
Collaborate with MLOps to productionize training/inference, including experiment tracking and model registry usage.
Write clean, testable code; review PRs and mentor other engineers on ML best practices.
Create technical docs for system design, model choices, and operational runbooks.

Required Qualifications :
Robust experience building NLP systems using PyTorch and Transformer-based architectures.




Hands-on experience implementing RAG pipelines end-to-end (retrieval, reranking, generation, evaluation).
Solid understanding of LLM behavior, prompting patterns, and failure modes (hallucination, tool misuse, leakage).
Experience with Git-based development, Docker, and CI/CD practices for ML services.
Working knowledge of relational and non-relational databases and how they fit into AI products.
Ability to design APIs/inference services that are reliable and maintainable

Preferred Qualifications
Experience with multimodal models and document understanding pipelines.
Experience with Kubernetes-based deployments and scaling strategies.
Knowledge of Azure ML and cloud-native patterns for production AI systems.
Experience with model distillation/quantization and serving optimization.
Prior experience mentoring or leading a small technical workstream.

Core Tools & Technologies :
PyTorch, Transformers; optional TensorFlow familiarity
RAG components: embeddings, vector stores, rerankers, eval harnesses
Docker; CI/CD; Git
Datastores: SQL + NoSQL; optional vector databases
Kubernetes; Azure compute & services (where applicable)

📌 Ai Engineer Nlp Chennai (India)
🏢 SrinSoft
📍 India