Location: Kolkata (On-site)
Experience: 4-5 Years
Education: Master’s degree in computer science /engineering/ data science
Employment Type: Full-Time

Role Overview

Additionally, S/HE should be a team player with solid communication skill and good command of English.

Key Responsibilities
- Build, train, and deploy advanced ML and AI models (supervised, unsupervised, deep learning)
- Design and implement LLM-powered AI agents and multi-agent workflows:
- Develop and optimize Retrieval-Augmented Generation (RAG) pipelines:
- Implement end-to-end ML lifecycle: data ingestion, feature engineering, training, deployment, monitoring
- Fine-tune models and optimize prompts for performance and cost
- Deploy production-grade AI systems using Docker, Kubernetes, and cloud platforms
- Monitor models for drift, bias, hallucinations, and performance degradation
- Collaborate with product and engineering teams to translate business problems into AI-driven solutions

Core ML & AI Expertise (Must have)

Machine Learning
- Regression, Classification (Logistic, Random Forest, XGBoost, LightGBM)




- Clustering (K-Means, DBSCAN)
- Time Series (ARIMA, Prophet, LSTM)
- Anomaly Detection (Isolation Forest, One-Class SVM)
- Recommender Systems

Deep Learning & NLP
- CNNs, RNNs, LSTM, Transformers
- BERT and Transformer-based architectures
- Embeddings and semantic search

Generative AI & Agentic Systems
- LLM API integration (OpenAI, Anthropic, AWS Bedrock)
- Prompt engineering (few-shot, chain-of-thought, structured prompting)
- LLM fine-tuning and evaluation
- Agent frameworks: LangChain, LlamaIndex, AutoGen, CrewAI
- Tool calling & function execution frameworks
- Vector databases (Pinecone, Weaviate, FAISS)
- Hybrid search (semantic + keyword)

Technical Stack
- Languages: Python (Primary), SQL
- ML Frameworks: Scikit-learn, TensorFlow, PyTorch
- API Frameworks: FastAPI / Flask
- MLOps: MLflow, CI/CD pipelines, model monitoring
- Cloud: AWS (Bedrock, SageMaker), Azure OpenAI, GCP AI
-

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