Senior Data Scientist/ AI Architect

Agentic AI, Context Layer & Knowledge Engineering

Location: Pune

Work Model: Hybrid- 3 days per week from office

Experience: 15-20 years

Employment Type: Full-time

About the Role:

We are looking for a strong, hands-on technical Data Science leader with 1520 years of experience to drive our next generation of AI initiatives, with a sharp focus on agentic AI systems, the context layer, and the data representation layer, Gemini adoption that grounds them in enterprise data. Prior experience at Google or deep, production-level expertise with the Google Cloud / Vertex AI ecosystem is strongly preferred.

You will be the senior-most technical authority on data science and applied AI in the organization: setting strategy, architecting agentic and ML systems end to end, and personally solving the hardest problems while mentoring a team of senior scientists.

Key Responsibilities:

Agentic AI:

- Lead the design, development, and production deployment of agentic AI systems autonomous and semi-autonomous agents that plan, reason, invoke tools, and execute multi-step business workflows.

- Architect multi-agent orchestration, tool-calling patterns, human-in-the-loop controls, guardrails, and agent evaluation frameworks.

- Define standards for agent reliability, safety, observability, and cost efficiency in production.

Context Layer:

- Own the architecture of the context layer that powers our LLM and agent applications: retrieval-augmented generation (RAG), vector search, embeddings, semantic caching, memory management, and knowledge graphs.

- Ensure agents and models are grounded in accurate, governed enterprise data with appropriate access controls, lineage, and freshness guarantees.

- Drive context engineering best practices chunking, retrieval quality, re-ranking, prompt/context optimization, and evaluation of groundedness and hallucination rates.

Data Representation Layer:

- Design and own the data representation layer of the AI platform: enterprise business glossary, ontologies, taxonomies, controlled vocabularies, and the semantic/metadata layer that gives shared meaning to data across the organization.

- Build and maintain knowledge graphs and entity models that connect business concepts, metrics, and data assets, enabling consistent interpretation by both humans and AI agents.

- Deploy AI agents to automate semantic curation agents that extract candidate glossary terms, propose taxonomy classifications, detect ontology drift, map source data to canonical concepts, and keep definitions, synonyms, and relationships up to date with human-in-the-loop review.

- Establish governance for semantic assets: term stewardship workflows, versioning, lineage, and alignment with data catalog and metadata management tools.

- Ensure the representation layer feeds directly into the context layer, so agents retrieve and reason over well-defined, unambiguous business concepts.

Core Data Science & Leadership:

- Define and own the overall data science and applied AI strategy and technical roadmap.

- Build and deploy large-scale machine learning and statistical models (prediction, ranking, recommendation, forecasting, NLP) that deliver measurable business impact.

- Design and interpret large-scale experiments (A/B testing, causal inference) to validate AI and product decisions.

- Translate ambiguous business problems into well-defined AI solutions and communicate results to C-level stakeholders.

- Mentor senior and staff-level data scientists; raise the bar on rigor, reproducibility, and engineering quality.

- Establish responsible AI, model governance, privacy, and compliance practices.

Required Qualifications:

- 1520 years of professional experience in data science, machine learning, or applied AI, with a consistent record of shipping production-grade systems at scale.

- Prior experience at Google (or an equivalent top-tier technology company), working on large-scale data, ML, or AI products.

- Hands-on experience building agentic AI solutions using frameworks such as LangGraph/LangChain, CrewAI, AutoGen, or Google's Agent Development Kit (ADK) / Vertex AI Agent Builder including tool calling, multi-agent orchestration, and agent evaluation.

- Proven experience designing the context layer for LLM applications: RAG pipelines, vector stores (Vertex AI Vector Search, Pinecone, FAISS, or similar), embedding models, context/prompt engineering, and enterprise data grounding.

- Hands-on experience building the data representation layer: business glossaries, taxonomies, ontologies (e.g., OWL/RDF/SKOS/Google Knowledge Catalog), knowledge graphs (e.g., Neo4j, graph databases, or GCP-native approaches), semantic modeling, and metadata/catalog integration (e.g., Dataplex, Collibra, Alation) including experience deploying agents or automation to build and maintain these semantic assets.

- Expert-level Python and SQL, with strong software engineering fundamentals (testing, CI/CD, code review, MLOps/LLMOps).

- Deep expertise across classical ML and deep learning: supervised/unsupervised learning, NLP, recommendation systems, forecasting, and ranking.

- Strong command of the GCP data and AI stack: BigQuery, Vertex AI, Dataflow, Dataproc, Pub/Sub, TensorFlow/JAX.

- Strong grounding in statistics, large-scale experimentation, and causal inference.

- Master's or PhD in Computer Science, Statistics, Mathematics, or a related quantitative field (or equivalent practical experience).

- Willingness to work from our Pune, Hyderabad, or Bengaluru office 3 days per week.

Preferred Qualifications:

- Direct experience with Google-internal ML/AI tooling and practices, or GCP Professional certifications (e.g., Professional Machine Learning Engineer).

- Familiarity with agent interoperability and context standards such as Model Context Protocol (MCP) and Agent2Agent (A2A).

- Experience with LLM fine-tuning, evaluation harnesses, and cost/latency optimization for agentic workloads.

- Publications, patents, or open-source contributions in ML, LLMs, or agentic AI.

- Experience building or scaling a data science / AI team from the ground up.

- Domain experience in e-commerce, fintech, adtech, or healthcare.

Onix is an equal opportunity employer. We celebrate diversity and are committed to creating an inclusive environment for all employees.