AVP — Data & ML Platform
Focus Areas: (i) Data Platform Engineering, (ii) ML Platform & MLOps, (iii) Platform Operations & FinOps, (iv) Data Governance & Quality
Location: Mumbai, India (Reliance Corporate Park / Global HQ) Grade: Assistant Vice President
Experience: 14–20 years total | 8–12 years in Data/ML Platform Engineering
Core Platform: Databricks Intelligence Platform (Unity Catalog, Delta Lake, MLflow, Mosaic AI)
The Context
Reliance Retail has completed Phase-1 of its group-wide transition to a unified Databricks Intelligence Platform. The foundation is set. We are now building the “v2.0” intelligence layer on top of this Lakehouse—standardising MLOps, scaling Agentic AI, and ensuring that the platform delivers sub-second latency for the full retail estate spanning several tens of thousands of stores and high-scale digital fronts.
The Data & ML Platforms group (Group A in Enterprise IT) is the engine room for this transformation. It is led by a VP (L2) and structured around 4 AVP-led pillars, 10 AI-ready Platform Engineers, and a transitioning cohort of Data Engineers. Each AVP owns a distinct platform layer and operates as a builder-leader—expected to architect, code-review, and ship alongside their team, not just manage.
The Four Pillars
We are hiring 4 AVPs, each leading one of the following platform pillars. While each AVP owns their pillar end-to-end, all four operate as a tight leadership team under the VP. Candidates may be considered across pillars based on strength and fit.
(i) Data Platform Engineering
Mission: Own the core Lakehouse infrastructure—the storage, compute, and developer platform layers that everything else runs on.
- Architect and operate the Delta Lake storage tier, Photon compute engine, and Unity Catalog abstraction layer for 1,000+ developers across retail verticals.
- Drive deep-tier optimisation: query plan tuning, cluster auto-scaling policies, Z-ordering strategies, and partition design for trillion-row datasets.
- Own the internal developer platform: SDKs, CLI tools, templates, and self-service onboarding that reduce time-to-first-query for new teams.
- Lead the technical remediation of Phase-1 migration debt—schema standardisation, pipeline consolidation, and SOR deduplication across hundreds of source systems.
- Manage the Data Engineer transition cohort assigned to this pillar, establishing engineering standards, code review practices, and career ladders.
(ii) ML Platform & MLOps
Mission: Industrialise ML—build the infrastructure that takes models from notebook to production at retail scale.
- Build and operate the end-to-end ML lifecycle using MLflow: experiment tracking, model registry, automated retraining, A/B testing, and canary deployments.
- Architect the real-time inference stack: model serving infrastructure for sub-100ms latency across recommendation, pricing, and demand forecasting use cases.
- Build the Agentic AI infrastructure: RAG pipelines, vector stores, fine-tuning workflows for Foundation Models (via Mosaic AI), and agent orchestration frameworks.
- Establish Feature Store governance: standardised feature definitions, freshness SLAs, lineage tracking, and cross-team feature reuse across retail verticals.
- Own ML platform reliability: GPU/TPU cluster management, training job scheduling, cost attribution per model, and incident response for production model degradation.
(iii) Platform Operations & FinOps
Mission: Keep the platform running, fast, and economically efficient—especially when it matters most.
- Own 99.99% platform availability, including war-room leadership during festive sales, store launches, and Reliance Retail peak events.
- Build and operate the FinOps practice: DBU cost attribution by team/workload, chargeback models, automated right-sizing, and executive cost dashboards.
- Architect the monitoring and observability stack: pipeline health, query performance, cluster utilisation, and data freshness SLAs across all 6 value streams.
- Drive capacity planning: forecast compute/storage demand against retail seasonality (festive cycles, new store openings, category launches) and pre-provision accordingly.
- Own incident management, runbooks, and post-mortem processes for the Databricks platform, ensuring mean-time-to-recovery targets are met and improved upon.
(iv) Data Governance & Quality
Mission: Be the technical guardian of India’s largest consumer dataset—make it trusted, compliant, and discoverable.
- Build “Governance-as-Code” systems on Unity Catalog: automated access policies, data classification, PII masking, and audit trails for DPDP Act compliance.
- Architect the data quality framework: automated profiling, anomaly detection, schema enforcement, and freshness monitoring across thousands of datasets.
- Own the data catalogue and discovery layer: metadata management, lineage visualisation, business glossary, and search interfaces that make data findable for 1,000+ users.
- Establish consent management infrastructure: track, enforce, and audit user consent signals across the full “Phygital” retail ecosystem (online + offline).
- Drive cross-pillar data standards: naming conventions, SOR deduplication rules, master data alignment, and data contract definitions between producer and consumer teams.
Minimum Qualifications (All Pillars)
- 14–20 years of professional experience in software engineering, data engineering, or ML infrastructure, with at least 3 years leading a platform team of 5+ engineers.
- 8–12 years of hands-on experience building and scaling data or ML platforms (e.g., Lakehouse architectures, Feature Stores, Streaming Engines, or MLOps pipelines).
- Strong technical depth in the Databricks ecosystem or equivalent distributed data platforms (Spark, Presto/Trino, Flink, or Kafka at scale). Databricks-specific experience is strongly preferred.
- Demonstrated “builder-leader” style: you still review code, debug production issues, and make architectural decisions—you haven’t fully delegated the technical details.
- Experience working in a large, complex technology organisation with inherited teams, cross-functional dependencies, and enterprise-grade compliance requirements.
- Bachelor’s or Master’s degree in Computer Science, Data Science, or a related field. Equivalent depth through industry experience and open-source contribution is equally valued.
Preferred Qualifications
- Prior experience at India-scale data platforms (multi-billion events/day, petabyte-class warehouses, or real-time serving at 10K+ QPS).
- Hands-on experience with MLflow, Mosaic AI, or equivalent ML infrastructure at production scale (not just experimentation).
- Familiarity with retail or e-commerce domain data: product catalogues, inventory systems, order management, customer signals, or supply chain datasets.
- Track record of building internal tooling or developer platforms that achieved organic adoption within a large engineering organisation.
- Experience with FinOps practices: DBU/compute cost attribution, chargeback models, or cloud cost optimisation at enterprise scale.
- Exposure to Indian data privacy frameworks (DPDP Act) or equivalent global regulations (GDPR, CCPA) in a data platform context.
Organisation Context
This role reports to the VP & Head of Data & ML Platforms, who reports to the Head of Enterprise IT, who reports to the CEO. You will be a peer of 3 other AVPs within the Data & ML Platforms group, and will work closely with 10+ AI-ready Platform Engineers (Architect & Principal levels) and the transitioning Data & Platforms Engineers’ cohort.
The broader Enterprise IT organisation includes 5 other L2 groups: CISO/Cybersecurity, HR/Finance/Legal Platforms, SAP-Core, Systems & AI Architects, and CIO + Cloud & Infrastructure. The retail portfolio spans all businesses including, but not limited to - AJIO, Tira, JioMart & Reliance Digital.