AI Engineer — LLM / VLM

SAI Group Ltd; Kolkata District; On-site; INR 1,500,000 - 3,500,000; Full time

Job summary

SAI Group Ltd is seeking an AI Engineer specializing in Large Language Models (LLMs) and Vision-Language Models (VLMs) to design, develop, and deploy production-grade AI solutions in Kolkata. The role focuses on prompt engineering, RAG, fine-tuning, multimodal AI, and model serving.

The candidate should have hands-on experience with LLMs/VLMs, Transformers, PyTorch, and Hugging Face, and be able to build production-grade APIs and pipelines using Python, FastAPI, Docker, and cloud platforms.

Qualifications

• Strong Python programming and software-engineering fundamentals.
• Hands-on experience with LLMs and/or VLMs.
• Strong understanding of Transformers, attention mechanisms, tokenization, embeddings, and inference.
• Experience with PyTorch and Hugging Face Transformers.
• Experience building RAG systems and vector-search solutions.
• Knowledge of prompt engineering and LLM evaluation.
• Experience with APIs, REST services, Git, Docker, and CI/CD.
• Familiarity with vector databases such as FAISS, Milvus, Pinecone, Weaviate, or pgvector.
• Understanding of cloud AI infrastructure, preferably AWS/Azure/GCP.

Responsibilities

• Develop and deploy AI applications using LLMs and VLMs.
• Build RAG pipelines involving document ingestion, chunking, embeddings, retrieval, reranking, and generation.
• Work with models such as GPT, Claude, Gemini, Llama, Mistral, Qwen, and multimodal/VLM models.
• Develop multimodal solutions involving text, images, PDFs, charts, tables, and documents.
• Perform prompt engineering, supervised fine-tuning, LoRA/QLoRA, and model evaluation.
• Build AI agents and tool-calling workflows where appropriate.
• Optimize inference for latency, throughput, memory, and cost.
• Develop APIs and production services using Python, FastAPI, Docker, and cloud platforms.
• Implement evaluation frameworks to measure accuracy, hallucination, relevance, latency, and safety.
• Collaborate with ML engineers, software engineers, and product teams to take prototypes into production.

Skills

Python programming

LLMs/VLMs experience

Transformers/attention

Prompt engineering

RAG pipelines

APIs REST

CI/CD

Cloud platforms (AWS/Azure/GCP)

Vector databases knowledge

Tools

Docker

Git

REST APIs

CI/CD tooling

PyTorch

HuggingFace Transformers

FastAPI

FAISS

Milvus

Pinecone

Weaviate

pgvector

Job description

Role Overview

We are looking for anAI Engineer specializing in Large Language Models (LLMs) and Vision-Language Models (VLMs)to design, develop, and deploy production-grade AI solutions. The ideal candidate should have strong experience with LLM/VLM architectures, prompt engineering, RAG, fine-tuning, multimodal AI, and model serving.

Key Responsibilities

• Develop and deploy AI applications using LLMs and VLMs.
• Build RAG pipelines involving document ingestion, chunking, embeddings, retrieval, reranking, and generation.
• Work with models such as GPT, Claude, Gemini, Llama, Mistral, Qwen, and multimodal/VLM models.
• Develop multimodal solutions involving text, images, PDFs, charts, tables, and documents.
• Perform prompt engineering, supervised fine-tuning, LoRA/QLoRA, and model evaluation.
• Build AI agents and tool-calling workflows where appropriate.
• Optimize inference for latency, throughput, memory, and cost.
• Develop APIs and production services using Python, FastAPI, Docker, and cloud platforms.
• Implement evaluation frameworks to measure accuracy, hallucination, relevance, latency, and safety.
• Collaborate with ML engineers, software engineers, and product teams to take prototypes into production.

Required Skills

• Strong Python programming and software-engineering fundamentals.
• Hands-on experience with LLMs and/or VLMs.
• Strong understanding of Transformers, attention mechanisms, tokenization, embeddings, and inference.
• Experience with PyTorch and Hugging Face Transformers.
• Experience building RAG systems and vector-search solutions.
• Knowledge of prompt engineering and LLM evaluation.
• Experience with APIs, REST services, Git, Docker, and CI/CD.
• Familiarity with vector databases such as FAISS, Milvus, Pinecone, Weaviate, or pgvector.
• Understanding of cloud AI infrastructure, preferably AWS/Azure/GCP.

VLM / Computer Vision Skills

• Experience with multimodal models such as Qwen-VL, LLaVA, Gemini, GPT vision models, or similar.
• Understanding of image preprocessing and document/image understanding.
• Experience with OCR, document intelligence, image classification, object detection, or visual question answeringis a plus.
• Ability to build pipelines combining vision + language + retrieval.