Role Overview

This role requires hands-on experience in LLM integrations, prompt engineering, vector databases, and production-grade deployment of AI solutions.

Experience Requirements
- 5-7 years of experience in Data Science / Machine Learning
- Minimum 1-2 years of hands-on experience with LLM-based systems:
- Experience building and deploying AI/ML solutions in production environments

Location & Employment Type
- Location: On-site - KOLKATA
- Employment Type: Full time

Must-Have Skills (Critical Requirements)

? LLM Integration
- Experience integrating LLM APIs such as:
- OpenAI:
- Anthropic:
- Amazon Web Services Bedrock
- Experience handling API orchestration, rate limits, token optimization, and cost control

? Agent Frameworks
- Hands-on experience with:
- LangChain:
- LlamaIndex:
- AutoGen:
- CrewAI
- Building multi-agent workflows
- Tool calling and function execution frameworks

? RAG Systems
- Designing and implementing Retrieval-Augmented Generation pipelines
- Experience with:
- Vector databases (Pinecone, Weaviate, FAISS, etc.)
- Embedding models




- Hybrid search (semantic + keyword)
- Chunking strategies and context optimization

? Model Fine-Tuning & Prompt Engineering
- Fine-tuning open-source LLMs
- Prompt design and evaluation
- Few-shot, chain-of-thought, and structured prompting
- Evaluation frameworks (accuracy, hallucination detection, grounding)

? ML Ops & Deployment
- Model deployment using Docker / Kubernetes
- CI/CD pipelines for ML systems
- Monitoring model performance and drift
- Logging, observability, and cost tracking
- Experience with cloud platforms (AWS / Azure / GCP)

Good-to-Have Skills (preferred)
- Experience with open-source LLMs (Llama, Mistral, etc.)
- Knowledge of RLHF / alignment techniques
- Knowledge graph integration
- Experience with real-time AI systems
- Experience in enterprise SaaS environments
- Exposure to Generative AI security & governance

Technical Stack

Languages:
- Python (Primary)
- SQL
- Optional:

📌 Data Scientist (Kolkata)
🏢 Ispeck Digital Solutions
📍 Kolkata