Check Out Our Innovative Solutions

Overall relevant experience: 3–4 years of overall relevant experience

Location: Pune

Work mode: Work from office

Role Summary

At Emergys, the Data Scientist – GenAI & Agentic AI will help turn business questions into practical, data-driven and AI-enabled solutions. The role combines core data science and predictive modelling with hands-on development of GenAI applications and agent components. Working with guidance from senior team members, the successful candidate will analyse data, build and test models and prototypes, and contribute to solutions that can be validated with stakeholders and developed further for production use.

Key Responsibilities

• Work with business and technical colleagues to clarify problem statements, understand available data, identify assumptions and define measurable success criteria.
• Prepare, explore and analyse structured and unstructured data using Python and SQL; assess data quality and communicate relevant findings.
• Build, test and validate analytical solutions and machine-learning models, including regression, classification, clustering, feature engineering and appropriate baseline comparisons.
• Apply statistics, experimentation and model-validation techniques to assess model performance, avoid data leakage and identify limitations.
• Contribute to GenAI applications using LLM APIs, prompt design, structured outputs, embeddings and retrieval-augmented generation (RAG), selecting techniques appropriate to the use case.
• Build and test agent components such as tool/function calls, API integrations and simple multi-step workflows under guidance; support workflow orchestration and state handling.
• Implement basic safeguards and failure handling, including input validation, retries, fallback responses, clear error handling and escalation to a human when needed.
• Develop rapid proof-of-concepts, demos and analytical artifacts; iterate based on feedback and document what has been validated, what remains uncertain and what should be tested next.
• Evaluate solutions using relevant measures such as predictive performance, groundedness, accuracy, reliability, explainability, latency, cost and business usefulness.
• Collaborate with senior data scientists and engineers to improve code quality, testing, version control and readiness for deployment; contribute reusable code, prompts, evaluation cases and other accelerators.
• Explain findings and technical choices clearly to teammates and stakeholders, using visualisations or demonstrations where useful.
• Delivers well-tested models, analyses and agent components against agreed requirements and timelines.
• Demonstrates appropriate model and approach selection, supported by validation and comparison with a reasonable baseline.
• Produces measurable prototype results against defined criteria, including quality, reliability, latency and cost where relevant.
• Identifies limitations and failure cases early and documents clear next steps for improvement.
• Contributes reusable components and communicates findings in a way that helps stakeholders make decisions.

Required Skills and Experience

• Strong working knowledge of Python for data analysis and application development, including readable, modular code; familiarity with common libraries such as pandas and NumPy.
• Working knowledge of SQL for querying, joining and analysing data.
• Solid foundations in statistics, exploratory data analysis, feature engineering, predictive modelling and model validation.
• Practical experience building and evaluating at least one or more machine-learning approaches such as regression, classification or clustering.
• Hands-on exposure to LLM-based applications and APIs, including prompt design and structured outputs; understanding of embeddings and RAG concepts.
• Foundational experience with agentic patterns such as tool/function calling, API integration and multi-step workflow execution, with a willingness to learn orchestration and state management.
• Ability to test outputs systematically and recognise failure cases, unsupported responses, data-quality issues and model limitations.
• Familiarity with Git and basic software engineering practices, including testing, debugging and documentation.
• Ability to communicate analytical findings and collaborate effectively with technical and non-technical colleagues.

Preferred Skills

• Exposure to an agent or orchestration framework such as LangGraph, LangChain, Semantic Kernel or an equivalent.
• Experience with a vector database or search platform, and with RAG evaluation or LLM evaluation tools such as RAGAS, DeepEval, Promptfoo, LangSmith or equivalent.
• Exposure to lightweight prototyping tools such as Streamlit, Gradio, notebooks or similar.
• Familiarity with cloud AI services, Docker, CI/CD, model deployment or MLOps/LLMOps concepts.
• Experience with NLP, transformer-based models, time-series analysis, optimisation or scenario modelling.
• Exposure to a business domain such as enterprise operations, manufacturing, supply chain, finance or customer service.

Education

• Bachelor’s or Master’s degree in Computer Science, Data Science, Industrial Engineering, Operations Research, Economics, or a related field is preferred.