Data Architect
Data Engineering | Hybrid in Islamabad, Punjab Province and Federal Capital Territory (ICT), Pakistan | Full Time
About the Role
Convo is looking for a Data Architecture Team Lead who will lead the data architecture team for a new data and decision platform that the company is building from the ground up for consumer-goods manufacturers. Reporting to the VP of Engineering, you will work closely with other leaders across Operations and Architecture to define the platform's technical direction and establish its foundational architecture.
The platform will ingest transaction, ERP, and third-party market data at scale, transform it into governed, contract-backed data products, and drive automated commercial decisions and settlements for customers operating across multiple countries and regulatory environments. Nothing is inherited. You will have complete ownership of designing the data model, data contracts, processing engine, storage and serving layers, tenancy model, and the team's ways of working.
This is a hands-on leadership role at the beginning of a major platform build. You will set the architecture, validate critical technical decisions through prototypes and performance measurements, recruit and lead the architects responsible for implementing the vision, and remain closely involved with the code to ensure the architecture works effectively in practice.The decisions made during the first year will shape the product for the next decade. We are looking for someone who has made these kinds of foundational architectural decisions before and has experienced their long-term implications.
Job Responsibilites:
Architecture
- Design the target data architecture from a blank page: canonical data model, data-contract standards, storage layering, processing engine, serving patterns, and the boundaries between ingestion, computation and consumption.
- Make the foundational technology choices and prove them: distributed versus in-process compute, batch versus streaming or incremental refresh, table formats, catalog and lineage, orchestration. Each choice is backed by a measured prototype and written up as an ADR.
- Design for many customers from day one: configuration-driven behaviour, explicit tenancy and isolation boundaries, versioning and upgrade paths that never require per-customer forks.
- Build in governance as engineering: contract enforcement at every layer, schema evolution rules, quality gates with clear blocking and tolerant modes, end-to-end lineage and provenance, auditability for financial outputs.
- Define the deployment model with security and DevOps: customer-cloud and vendor-hosted options, identity-based access, least privilege per pipeline, secrets handling, infrastructure as code.
- Set the non-functional targets (latency, throughput, cost per customer, recovery) and the way they are measured, before the first production customer.
Leadership
- Build and lead the data architecture team: define the roles, hire the first architects and senior engineers, set objectives, coach, and run design reviews.
- Work day to day with the Operations and Architecture leads as a peer group, and be the technical counterpart to product management, the AI and data-science team and go-to-market; translate commercial and regulatory requirements into architectural constraints and back.
- Establish the architecture decision process for the product: a lightweight ADR practice, a review forum, and a visible record of what was decided and why.
- Mentor engineers on pipeline design, performance and data modelling; set the definition of done for a data product.
- Represent the data architecture to early customers, partners and auditors.
Requirements:
- We care about what you have designed, shipped and operated, not which language or vendor you did it in. Candidates from JVM, Python, SQL-first, Scala, Rust or Go backgrounds are equally welcome.
- 10+ years building large-scale data pipelines, with at least 3 of them in a lead or principal architect role accountable for the design across a team.
- Designed and operated at least two production platforms at terabyte-plus scale that ingested from heterogeneous external sources (ERPs, third-party data providers, files, APIs) and served downstream applications with SLAs. You can describe what broke, what you changed and what it cost.
- Fluency with distributed data processing on at least one major engine (Spark, Flink, Trino, BigQuery, Snowflake, or equivalent) and with in-process alternatives (DataFrame libraries, columnar formats), and a clear view of when each is the wrong choice.
- Data modelling across styles: dimensional, medallion/lakehouse, data vault, or domain-driven; schema evolution and contract-first design in practice, not just on slides.
- Data quality, lineage and governance as engineering disciplines: contract enforcement, quarantine strategies, provenance, auditability, and the trade-offs between blocking and tolerant validation.
- Orchestration and reliability: dependency-aware workflow engines, idempotent and resumable jobs, backfills, incremental refresh, cost control; experience running pipelines in production, including on-call and incident review.
- Cloud-native architecture on at least one hyperscaler, including identity-based access, secrets management, container orchestration and infrastructure as code; able to reason about a second cloud even if you have not shipped on it.
- Multi-client or multi-tenant platform design: configuration-driven behaviour, tenancy boundaries, upgrade and versioning strategies for a product used by many customers.
- Leadership evidence: hired and grown engineers or architects, run design reviews, and made decisions that held up under disagreement. You write clearly; ADRs, design docs and review comments are part of how you lead.
- Judgement under commercial pressure: you can explain a technical trade-off to a product owner or a client and hold a line when the short-term ask is wrong.
- Fluent professional English.
Nice to Have
Consumer-goods, retail or revenue-management domain exposure: trade promotion, sell-in/sell-out/sell-through data, syndicated panel data, or ERP master data.
Experience feeding machine-learning pipelines and serving model outputs through APIs; familiarity with feature stores and model registries.
Open data-contract standards (ODCS or similar) and data-quality frameworks such as Great Expectations or Soda.
Lakehouse table formats (Delta, Iceberg, Hudi) and catalog or lineage tooling (Unity Catalog, OpenLineage, DataHub).
Observability for data systems: OpenTelemetry, metrics design, cardinality discipline.
Experience taking a platform from first commit to its first paying customers, and the scars to show for it.
What we have for you:
- Great compensation package, medical benefit for you and your family, free lunch, annual performance-tied increments & performance recognition awards and a great lean and agile work culture!
- Convo endorses a culture of diversity in all aspects and aims to build a diverse team of amazing individuals!
