Mansukh
Sonawala

Building the financial data nervous system for one of the world's largest sneaker and collectibles marketplaces — every order, payment, fee, and tax event flows through pipelines I own into audit-ready ledger entries used for global financial close.

• Own the end-to-end subledger-to-GL pipeline: ingesting raw transactional events from Kafka and the BlackLine API, processing them in Databricks using Scala and Spark, and mapping outputs to General Ledger accounts across multiple regions and accounting systems.
• Designed scalable batch and real-time data pipelines on Databricks processing high-volume order, payment, fee, and tax transactions — built with strong consistency guarantees and full audit traceability.
• Built automated data quality and reconciliation frameworks that catch ledger discrepancies before they reach Finance, significantly reducing downstream reporting incidents.
• Partner with Accounting, Finance, and Tax teams to support period-end close, trial balance validation, and SOX-compliant reporting across multiple global regions.
• Delivered Tableau dashboards giving Finance real-time visibility into revenue, taxes, account balances, and reconciliation status — used in board-level reporting.
• Owned infrastructure automation via Terraform and AWS; managed complex orchestration with Airflow and Databricks Workflows; acted as technical owner driving architecture decisions across the financial data domain.
• Integrated Claude Code and AI-assisted workflows into daily engineering — using LLMs for SQL optimization, root-cause analysis, code generation, and documentation, measurably accelerating delivery cycles.

Took ownership of the analytics platform at a fast-moving PropTech startup — stabilised unreliable pipelines, migrated the entire BI layer, and built a data service the engineering team still depends on.

• Designed and orchestrated 15+ production ETL pipelines using Airflow, Airbyte, AWS Glue, and dbt — improving reliability and giving analytics teams clean, timely data to work with.
• Built a FastAPI-based data service with structured logging and PagerDuty alerting, dramatically cutting incident response time and reducing outages that previously interrupted business reporting.
• Led the migration from Tableau to Looker — a full BI platform switch covering data model re-architecture, report recreation, and stakeholder rollout. Improved query performance and reduced platform costs.
• Owned deployment pipelines via AWS Fargate ECS and GitHub Actions, and migrated Postgres databases to Multi-AZ clusters for better scalability and zero-downtime operations.
• Implemented centralised observability with CloudWatch and Grafana — the team went from discovering data issues at stand-up to getting alerted automatically.

High-throughput pipelines for an e-commerce roll-up managing dozens of brands — real-time Kafka streaming at 50M+ records/day, a Kubernetes-native data platform, and engineering team leadership across a 5-person group.

• Designed and automated 20+ production ETL pipelines, eliminating manual data processing and improving availability of business-critical datasets across multiple brands.
• Led batch and streaming pipelines processing 50M+ records per day using PySpark, Kafka, EMR, and dbt — maintaining 99.9% data availability on the streaming layer.
• Owned Airflow and Airbyte deployments on Kubernetes end-to-end — configuration, scaling, and on-call — while driving cost reduction through right-sized cluster management.
• Built SQS-based ingestion pipelines for marketing and third-party data, adding reliable delivery guarantees that batch jobs previously lacked.
• Established CI/CD pipelines with Kubernetes and Terraform, and led DevOps initiatives on AWS Fargate ECS that significantly cut deployment time.
• Led a 5-person cross-functional team — setting code standards, running reviews, and mentoring junior engineers while aligning delivery with business stakeholders across multiple brands.

Started my data engineering career building real-time Spark pipelines for Airbus — aircraft engine telemetry with hard latency requirements and zero tolerance for data loss.

• Developed and operated real-time Spark pipelines for Airbus, reducing latency in aircraft engine telemetry processing where data freshness had direct operational impact.
• Implemented AWS infrastructure using Terraform and CloudFormation — API Gateway, CloudFront, DynamoDB — building repeatable, auditable deployments from the start of my career.
• Built back-end services in Python on AWS, leveraging PySpark for large-scale data processing and throughput optimisation across high-volume telemetry datasets.
• Contributed to internal open-source tooling, establishing reusable components and engineering standards adopted across the team.