Job Title:  T&T - Customer - M&C - Senior Consultant | ML/AI/Data Science | Pune |

T&T - Customer - M&C - Senior Consultant | ML/AI/Data Science | Pune |
• Job requisition ID : 117129 
• Location: Pune
• Entity: Deloitte Touche Tohmatsu India LLP 

The team:

                                                              

Customer has to do much more than keep the wheels turning; it is the engine that drives functional excellence and the enabler of innovation and long-term growth. Learn more about: Customer

  

Key Responsibilities:

 

1. Data Pipeline Development & Maintenance

  • Design, build, and maintain scalable batch and near real-time data pipelines optimized for data science consumption patterns.
  • Develop robust ETL/ELT workflows using Snowflake Streams, Tasks, and Dynamic Tables to ingest and transform data from multiple sources.
  • Ensure pipelines deliver low-latency, high-freshness data feeds for model training, feature engineering, and reporting.
  • Automate pipeline workflows and implement orchestration using Airflow or equivalent tools with full lineage tracking.
  • Monitor and alert on pipeline health, SLA breaches, and data drift to proactively resolve issues before they impact downstream teams.

 

2. Snowflake Data Platform Engineering

 

  • Serve as the primary Snowflake platform engineer — owning schema design, warehouse sizing, resource monitors, and cost governance.
  • Design and implement dimensional models (star/snowflake schemas) and analytics-ready data layers in Snowflake to support both BI and ML use cases.
  • Build and maintain curated Gold/Semantic layers in Snowflake using SQL-based transformation logic, ensuring models are tested, documented, and version-controlled.
  • Leverage Snowflake-native features including Dynamic Tables, Snowpark (Python), Materialized Views, and Search Optimization for ML and analytics workloads.
  • Implement Snowflake data sharing, secure views, and row-access policies to enable governed, compliant data distribution across teams.
  • Optimize query performance through query profiling, clustering keys, partitioning strategies, and virtual warehouse auto-scaling.
  • Manage Snowflake environments across dev/staging/prod, enforcing RBAC, network policies, and data masking for sensitive fields.
  • Evaluate and adopt new Snowflake capabilities (e.g., Cortex AI/ML Functions, Apache Iceberg integration) to improve platform maturity.

 

3. SAS to Snowflake Migration & Modernisation

 

  • Lead the end-to-end migration of legacy SAS-based data pipelines to a modern Snowflake-native architecture — translating SAS DATA steps, PROCs, and macros into performant SQL and Python-based ELT workflows.
  • Re-engineer SAS transformation logic into Snowflake Tasks and Streams, preserving business rules and output fidelity while significantly improving execution speed, scalability, and maintainability.
  • Establish a migration validation framework to reconcile SAS outputs against Snowflake equivalents row-by-row — ensuring zero regression in data quality and giving business stakeholders confidence to sign off on cutover.

 

4. Data Science & Analytics Enablement (Primary Focus)

 

  • Act as the dedicated data provider for the Data Science team — understanding their feature requirements and translating them into reliable, performant datasets.
  • Build and maintain feature stores and curated feature tables in Snowflake to support reproducible ML model training and inference pipelines.
  • Design experiment-ready datasets with proper train/test splits, time-based partitioning, and versioning to ensure ML model integrity.
  • Collaborate with data scientists to understand analytical use cases and pre-aggregate or denormalize data to reduce query complexity for notebooks and dashboards.
  • Provide clean, structured, and high-quality data for statistical analysis, predictive modeling, and reporting — proactively flagging data anomalies.
  • Assist in feature engineering pipelines by building reusable transformation logic in Snowpark or Snowflake Tasks that data scientists can directly consume.
  • Support A/B testing and experimentation infrastructure by ensuring variant-split data is cleanly captured, versioned, and accessible.

 

5. Data Quality & Governance

 

  • Implement end-to-end data validation checks using Great Expectations or custom Snowflake-native validation frameworks.
  • Enforce data accuracy, completeness, freshness, and consistency SLAs across all pipelines serving data science and analytics consumers.
  • Maintain data contracts between source systems and downstream consumers, documenting schemas, update frequencies, and known limitations.
  • Work with stakeholders to define and enforce data standards, naming conventions, and governance best practices across the Snowflake platform.
  • Build and maintain a data catalog and lineage documentation so data scientists can self-serve and understand data provenance.

 

6. Performance Optimization & Cost Efficiency

 

  • Continuously profile and tune Snowflake queries used by data science notebooks, dashboards, and scheduled jobs.
  • Identify and resolve pipeline bottlenecks, redundant materializations, and inefficient storage patterns to reduce Snowflake compute costs.
  • Refactor legacy pipelines for scalability, applying modern ELT patterns and incremental loading strategies.
  • Implement table clustering, partitioning, and micro-partition management strategies to optimize large-scale analytical queries.

 

7. Collaboration & Stakeholder Management

 

  • Work cross-functionally with Data Scientists, Analysts, Product, and Engineering teams to translate business questions into data solutions.
  • Participate in sprint planning and backlog grooming with the data science team to prioritize data needs alongside engineering work.
  • Communicate effectively with both technical and non-technical stakeholders — explaining trade-offs, timelines, and data limitations clearly.

 

 

Desired Qualifications:

 

  • Base location: Pune 
  • Education: Professional Qualification - B.E./B.Tech/MCA/MBA/MS
  • This profile involves occasional travelling to client locations.
  • Hybrid is our default way of working. Each domain has customized the hybrid approach to their unique needs.