Data Scientist
Total positions:2
Positions Level
Sr. Consultant
Location: USI Location (Hyderabad, Bangalore, Pune, Mumbai, Kolkata, Chennai, Gurgaon)
Experience: 3-8 + Years
Shift Time: 11 am to 8 pm
Work Mode: Hybrid
Mandatory Skills:
Pyspark, Machine learning, Azure, ML Models, Azure, synapse, Data Bricker’s. phyton-Scripting, Lag chain
Role Summary
We are seeking a Data Architect to own canonical data modeling and abstraction discipline across our platforms, most notably our payroll-system-agnostic platform built on a Common Data Model (CDM) that normalizes multiple vendor payroll systems into one canonical structure. This role is distinct from a Data Integration Architect: rather than owning pipeline mechanics (ETL/ELT, lakehouse, vector stores), you own the correctness of the model itself - how source-system concepts get abstracted into stable canonical entities, how configuration (not hardcoding) drives extensibility as new sources and clients are onboarded, and how the platform's engineering stays reconciled against ground-truth data rather than assumptions. You will work closely with data engineering, application engineering, and delivery leadership to ensure the data model can scale to new payroll vendors and clients without being redesigned each time.
Experience bands:
- Manager: 12+ years total experience in data architecture, data modeling, or related engineering roles, including experience owning a canonical/common data model used across multiple clients or source systems
Key Responsibilities
- Own canonical data modeling: define stable, source-agnostic entities and relationships that represent the business domain (e.g., payroll constructs such as Gross Pay, Net Pay, Regular Pay, Bonus, Termination) independent of any single vendor's data shape.
- Design config-driven architecture: ensure new source systems, clients, or field variations are onboarded through configuration and mapping rules rather than one-off code changes or schema forks.
- Enforce correctness and reconciliation discipline: define and validate that canonical data reconciles against source system totals and known ground truth, and build in checks that surface drift or silent data-quality regressions early.
- Own production engineering and abstraction discipline: review and approve how abstraction layers are actually implemented in code, ensuring the conceptual model and the production implementation do not diverge over time.
- Partner with the Solution Architect and Data Integration Architect: this role owns the correctness and stability of the canonical model itself, while those roles own end-to-end platform coherence and pipeline/ingestion mechanics respectively.
- Lead data modeling reviews and design walkthroughs for new entities, new client onboarding, or new source-system integrations, pressure-testing proposed models against edge cases and real production data.
- Define data governance practices for the canonical layer: versioning of the model, change management for schema evolution, lineage, and documentation standards.
- Validate architecture and modeling decisions directly against real client data (not sampled or synthetic data) before sign-off, and flag contradictions between assumed and actual data behavior.
- Create architecture artifacts (canonical model documentation, ADRs, entity-relationship diagrams, mapping specifications) and maintain them as the model evolves.
- Mentor data engineers on modeling discipline and abstraction practices; review code and configuration for adherence to the canonical model rather than workarounds.
- Support onboarding of new foundation clients by assessing how their source systems map to the existing canonical model and where the model needs to be extended versus where a client-specific exception is being incorrectly introduced.
Required Qualifications
- 7+ years (Senior Consultant) or 11+ years (Manager) total experience in data architecture or data modeling roles, including demonstrated ownership of a canonical or common data model spanning multiple source systems or clients.
- Strong hands-on experience with dimensional and canonical data modeling, including designing entities that stay stable while underlying source systems vary or change.
- Demonstrated experience designing config-driven (not hardcoded) architecture for onboarding new data sources, clients, or variations without redesigning the core model.
- Strong track record of building or enforcing reconciliation and data-quality validation practices against ground-truth data at production scale.
- Proficiency in SQL and at least one programming language (Python preferred) sufficient to review and validate production data pipeline code against the intended model.
- Experience with relational databases (PostgreSQL preferred) for canonical model implementation at scale.
- Ability to distinguish role-level architectural responsibility from years of experience alone - evaluating what a candidate actually owned versus what they were exposed to or supervised.
- Hands-on proficiency with AI coding assistants (Claude Code, GitHub Copilot, Cursor, Windsurf, or equivalent) for architecture work, design documentation, and engineering workflows. Familiarity with agentic engineering patterns is expected.
- Strong communication and stakeholder management skills; ability to explain modeling trade-offs to both engineers and business/delivery stakeholders.
Preferred Qualifications
- Experience building a canonical data model specifically for payroll, HR, finance, or other transaction-critical domains with strict correctness requirements.
- Experience onboarding multiple vendor systems (e.g., ADP, Workday, SAP, or equivalents in other domains) into a single normalized model.
- Familiarity with orchestration tools (Prefect, Airflow, Dagster, or equivalent) sufficient to understand how the canonical model is populated and refreshed in production, even if not owning the orchestration layer directly.
- Experience with cloud data platforms (Azure preferred) hosting the canonical model and its supporting infrastructure.
- Experience defining data contracts or schemas consumed by downstream AI/ML models, understanding how model-readiness requirements should shape canonical model design.
- Experience in regulated industries and implementing audit-ready data governance and change-control practices.
- Cross-functional fluency with data integration, cloud, and AI architecture patterns - working familiarity with ETL/ELT and lakehouse concepts, cloud-native deployment, and LLM data consumption patterns. Not expected to own these, but must collaborate effectively with Data Integration, Cloud Solution, and AI Solutions architects in joint reviews.
Key Competencies
- Rigorous, first-principles approach to canonical modeling - able to separate what is a stable business concept from what is an artifact of a particular source system.
- Discipline around config-driven extensibility; instinctively pushes back on hardcoded, one-off solutions.
- Correctness and reconciliation mindset - treats data validation against ground truth as non-negotiable, not an afterthought.
- Production engineering ownership - reviews how abstractions are actually implemented, not just how they are designed on paper.
- Strong architectural judgment and trade-off management; comfortable pushing back and flagging contradictions rather than deferring to assumptions.
- Mentorship and technical leadership; comfortable with hands-on review of code, configuration, and data when needed.