Senior Data Scientist – GenAI & Agentic AI

• *Experience:** 9–10 years of overall relevant experience

• *Location:** Pune

• *Work mode:** Work from office

• *Role Summary**

At Emergys, the Lead Data Scientist – GenAI \& Agentic AI leads discovery, solution architecture and technical delivery for business\-critical data science, GenAI and agentic AI initiatives. The role brings structure to ambiguous business challenges, guides the choice of analytical and AI approaches, and sets the quality bar for rapid prototyping and production\-ready delivery. The Lead works directly with clients and cross\-functional teams, manages technical risks and trade\-offs, and helps convert successful prototypes into scalable solutions and reusable offerings.

• *Key Responsibilities**

• Lead discovery with clients and business stakeholders: clarify desired outcomes, frame loosely defined problems, identify constraints and risks, assess data readiness, and define success measures and a delivery roadmap.
• Own solution architecture and technical direction across data science, predictive modelling, GenAI and Agentic AI; explain why a selected approach is appropriate for the business problem.
• Guide teams in applying statistics, machine learning, experimentation, optimisation and LLM\-based methods, ensuring analytical rigour and fit\-for\-purpose complexity.
• Oversee the design and delivery of end\-to\-end solutions, from data exploration and POCs through evaluation, architecture decisions, production transition and operational readiness.
• Set technical patterns for agentic systems, including tool/function calling, API integration, workflow orchestration, state and memory management, multi\-step execution, permissions and controlled tool access.
• Define production standards for failure handling and resilience, including timeouts, retries, fallbacks, idempotency where applicable, escalation, human approval checkpoints and safe termination paths.
• Establish evaluation and quality standards for predictive and GenAI/agent solutions, covering business impact, task success, accuracy, retrieval quality, groundedness, explainability, reliability, latency, cost and risk.
• Guide production delivery with engineering and platform teams, including security and privacy controls, prompt\-injection mitigation, access controls, observability, monitoring, incident response, deployment and rollback considerations.
• Manage technical delivery risks, dependencies, scope, trade\-offs and stakeholder expectations; surface issues early and recommend pragmatic mitigation plans.
• Lead client discussions and technical readouts, communicating findings, prototype evidence, limitations, architecture choices and business implications to both executive and technical audiences.
• Create and promote reusable components, reference architectures, evaluation harnesses and accelerators that improve repeatability, delivery speed and productization potential.
• Set standards for code quality, testing, documentation, peer review and responsible AI; mentor senior and junior team members and raise overall technical capability.
• Partner with product, engineering and business teams to identify which validated prototypes should be scaled, productized or stopped, and define the transition path for selected solutions.
• Establish a clear problem definition, architecture and delivery path tied to measurable business outcomes.
• Guide teams to select the simplest effective approach and validates it against transparent technical and business criteria.
• Deliver prototypes and production solutions with evidence of quality, reliability, explainability, latency, cost and business impact.
• Anticipate and manages architecture, security, operational and delivery risks; ensures human approval and control points where appropriate.
• Improve delivery leverage through reusable accelerators, consistent evaluation standards, technical mentoring and successful productization of validated solutions.

• *Required Skills and Experience**

• 9–10 years of overall relevant experience in data science, applied AI, machine learning or related solution delivery, including substantial hands\-on technical work.
• Strong Python and SQL skills and sound software engineering judgement for designing maintainable data and AI solutions.
• Deep foundations in statistics, predictive modelling, feature engineering, model selection, validation and interpretation; ability to assess regression, classification, clustering and other methods as appropriate.
• Demonstrated experience leading discovery and translating ambiguous business problems into solution architecture, delivery plans and measurable outcomes.
• Hands\-on experience with GenAI applications and Agentic AI, including LLM APIs, tool/function calling, API integration, workflow orchestration and state management.
• Production delivery experience for AI/LLM or agentic systems, including failure handling, monitoring/observability, security controls and operational support.
• Experience designing evaluation strategies for models and agentic workflows, including task success, tool behaviour, reliability, explainability, latency, cost and business value.
• Strong judgement on architecture trade\-offs, build\-versus\-use decisions, complexity, scalability and cost optimisation.
• Experience leading technical delivery, managing risks and dependencies, and communicating decisions to clients and cross\-functional stakeholders.
• Ability to establish engineering and responsible\-AI standards, guide technical reviews and mentor colleagues across experience levels.

• *Preferred Skills**

• Experience with production agent frameworks such as LangGraph, LangChain, Semantic Kernel or equivalent; no single framework is mandatory.
• Familiarity with evaluation and observability tools such as LangSmith, RAGAS, DeepEval, Arize Phoenix, Promptfoo or equivalent.
• Experience with scikit\-learn and relevant deep\-learning frameworks such as PyTorch or TensorFlow, plus vector search/database technologies.
• Experience with cloud AI platforms, Docker, CI/CD, FastAPI or equivalent API frameworks, and MLOps/LLMOps.
• Experience with multi\-agent workflows, model/provider abstraction, asynchronous execution or event\-driven architectures.
• Experience designing human\-in\-the\-loop approvals, secure tool execution, privacy safeguards, enterprise governance and prompt\-injection mitigation.
• Experience in consulting, client services, enterprise product development or domains such as manufacturing, supply chain, finance and enterprise operations.

• *Education**

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