Data Architect

Location: Kochi, Delhi NCR (Full-time, On-site)

Company: Intelo.ai

About Intelo

Intelo is a transformation partner building AI agents that drive efficiency, accuracy, and growth across businesses. We solve critical retail and supply chain workflows using purpose-built AI technology.

Our platform enables retailers to:

- Forecast demand accurately
- Reduce excess inventory
- Improve sell-through
- Make faster, smarter decisions

We operate at the intersection of AI, data engineering, and product thinking - building systems that translate complex data into real business outcomes.

Role Overview

We are looking for a Data Architect to own the data architecture of our platform end to end - from how a retailer's raw feeds land, through our lakehouse layers, to the models our forecasting, allocation, replenishment, and planning agents read.

This is the role that decides how one platform serves many retailers without becoming many platforms. Every client arrives with a different product classification, a different fiscal calendar, different source systems, and different definitions of the same metric. Our answer has to be configuration and a well-designed model - not a fork per customer. You will own that model, the contracts around it, and the standards the rest of engineering builds against.

It is a hands-on architecture role. You will write SQL and Python, review pipelines, and be in the data when something does not reconcile. You will work alongside our architecture group and closely with the data engineering, integrations, and implementation teams.

What We Expect from Every Engineer at Intelo

This is part of our operating philosophy.

Here to Make a Difference

- Challenges the status quo
- Experiments boldly
- Adapts when the ground shifts

Open by Default
- Raises concerns early
- Shares knowledge transparently

Owns It
- Takes initiative
- Follows process even when it's hard
- Makes decisions with context - not escalation

Product Over Tickets
- Thinks from the customer backwards
- Fights unnecessary complexity
- Automates repetitive work

Always Learning
- Seeks deep understanding
- Uses AI as leverage to improve thinking and output
- Continuously upgrades technical and domain judgment

Key Responsibilities Multi-Tenant Data Modelling
- Own the canonical data model across product, location, sales, inventory, and planning - one model that absorbs every client's variation
- Design hierarchies as configuration, not code: retailers arrive with three-level, four-level, and industry-standard classification schemes, and all of them must resolve through the same model
- Model fiscal calendars, channels, and organisational structures as per-tenant configuration with sane platform defaults
- Draw the line between what is genuinely tenant-specific and what belongs in the platform - and hold that line





Lakehouse Architecture
- Own the medallion architecture (bronze / silver / gold) on Databricks and Delta Lake, including Unity Catalog structure and governance
- Design partitioning, clustering, and incremental/CDC strategies that keep query cost and scan volume under control as clients grow
- Define how versioned and layered data is stored and read - planning data in particular carries baselines, seeded layers, actuals, and plan versions that must resolve deterministically
- Set the standards for how new pipelines are built, reviewed, and registered

Data Contracts & Trust
- Define per-source data contracts: schema, cadence, delimiters, encoding, keys, and expected volumes agreed in writing before a pipeline is built
- Make missing data loud. A pipeline must not be able to report success while an expected source is absent - freshness SLAs per tenant per source, asserted rather than displayed
- Build reconciliation into the architecture: loaded data ties back to client control totals, and metrics can be traced from what a planner sees to the source file it came from
- Own data quality strategy - profiling, contract tests, and volume-delta checks that catch silent regressions

Serving & Semantic Layer
- Design the serving path from the lakehouse into the operational store and the query layer our agents and reports read
- Own the shared semantic layer so a metric is defined once and consumed identically by every worker, agent, and report
- Keep the contract between the lakehouse and the application stable while the internals evolve

Performance & Cost
- Own the cost and performance profile of the data platform: warehouse sizing, runtime and compute choices, job scheduling, and query design
- Diagnose and fix the expensive class of problems - misaligned partitions, unbounded scans, missing statistics - and prevent their recurrence
- Make cost per client visible and defensible

Client Onboarding Architecture
- Own the reference architecture for onboarding a recent retailer: source discovery, mapping to the canonical model, key strategy across multiple source systems, and attribute profiling
- Make the judgement calls that decide an implementation - global versus per-market product master, planning hierarchy and forecast grain, how to handle sparse or dirty attributes
- Reduce onboarding from bespoke engineering toward repeatable configuration

Governance, Security & Standards




- Own tenant isolation and access control in the lakehouse; support our SOC 2 posture with least-privilege access and auditability
- Write and maintain the reference documentation, patterns, and decision records the team builds against
- Review designs and pipelines, and raise the data engineering bar across the team through mentoring rather than gatekeeping

Required Qualifications
- 8+ years in data engineering or data architecture, with 3+ years owning architecture for a platform or product (not a single warehouse)
- Deep data modelling skill - dimensional modelling, slowly changing dimensions, hierarchies, and the judgement to keep a model simple
- Strong hands-on Spark and SQL; production experience with a lakehouse or warehouse platform (Databricks / Delta preferred; Snowflake or BigQuery acceptable)
- Demonstrated ownership of partitioning, incremental loading, and query performance at meaningful scale
- Strong Python; comfortable reading and reviewing application code
- Relational database depth (PostgreSQL preferred), including schema evolution and migration discipline
- Practical experience with data quality, contracts, or reconciliation - not just pipeline delivery
- Ability to explain a design decision and its trade-offs to engineers, product, and customers
- Bachelor's degree in Computer Science or related field, or equivalent practical experience

Preferred Experience

We would prefer architects with experience in:
- Retail, supply chain, or demand planning data - product hierarchies, fiscal calendars, assortment, and inventory
- Multi-tenant SaaS data architecture where clients differ structurally, not just by row
- Databricks Unity Catalog, Delta CDF, and Delta performance tuning
- Semantic or metric layers, and query-generation frameworks
- Azure data services and infrastructure as code (Terraform)
- Databricks cost optimisation
- Supporting SOC 2 or similar compliance in a data platform

What This Role Owns at Intelo A Data Architect at Intelo:
- Owns the model every agent, report, and integration depends on
- Turns client variation into configuration instead of forks
- Designs so that bad data is loud, not silent
- Treats cost and correctness as architectural properties, not afterthoughts
- Makes the next client onboarding faster than the last one
- Raises the data engineering bar for the whole team

Why Join Intelo
- Own the data architecture of a platform serving enterprise retail brands, at a stage where your decisions still shape it
- Real scale and real complexity - hundreds of millions of rows, many tenants, genuinely different data models
- Direct line to the business outcome your model produces
- High ownership and visibility, working alongside a small senior architecture group
- Fast-moving, product-first environment
- Culture that rewards initiative and deep thinking

📌 Data Architect (Delhi)
🏢 Intelo.ai
📍 Delhi