Salayhin Md Sirajus | Data Platform Architect | Solution Architecture & Data Engineering

Experience

Staff Data Platform Engineer

Zeals Co., Ltd · Senior Data & Infrastructure Engineer → Senior Data Platform Engineer → Staff Data Platform Engineer (promoted Nov 2023)

Tokyo, Japan (joined remotely, relocated 2022) Apr 2021 - Present

Current

Sole architect and primary technical decision-maker for the company-wide data platform used by analytics, ML, and BI teams. Operate the BI layer (Tableau, Redash) as platform-managed services.

Translate stakeholder requirements into technical specifications. Author design documents (HLD-style proposals, architecture decision records) and platform standards, and lead architecture and design reviews.

Defined and delivered the target architecture: a production Apache Iceberg lakehouse on GCS (BigLake Metastore), queried by both BigQuery and Spark on Dataproc. Selected based on documented trade-offs: cost, multi-engine reads without duplication, and protobuf schema evolution.

Built the team's core pipeline framework: a Docker- and script-driven system that provisions ephemeral Dataproc clusters. It runs Scala/Spark ingestion of approximately 100 GB/day (peaking at 1 TB) of protobuf events from Pub/Sub through landing, bronze, and gold medallion layers, with automated Iceberg table maintenance. Orchestrated end-to-end on Cloud Composer (Airflow), the platform's primary orchestrator.

Led adoption of dbt Cloud as the platform's modeling and semantic/metrics layer, with automated testing and documentation standards; analytics engineers now build and own models. Reduced stakeholder-facing data incidents from 2–3 per month to fewer than one every two months, and pipeline maintenance from 1–2 days/week to a few hours.

Own technical delivery practices: CI/CD via GitHub Actions for all platform workloads (Spark pipelines and dbt models), release readiness and UAT coordination with stakeholders, and runbooks for support readiness. Defined data contracts, quality standards, and SLAs across teams.

Own security and compliance by design: column-level PII masking at ingest (BigQuery policies), GDPR right-to-erasure across Iceberg tables and snapshots, and cloud resources, IAM, and secrets managed as Infrastructure-as-Code via Terraform with PR review. End-to-end data lineage and metadata via dbt and GCP-native lineage tracking; partnered with security engineers on SOC 2 compliance for the analytics layer.

Drove a platform-wide cost program spanning BigQuery optimization and the Iceberg migration, lowering monthly spend by approximately 60% (from ¥1.2M to under ¥500K).

Optimized BigQuery workloads using system tables and billing exports: partitioning, clustering, and gold-layer star-schema modeling (dim_/fact_ tables). Reduced query runtime by up to 90% and scan volume by 80%.

Delivered an end-to-end FinOps & observability application on GKE (Kubernetes): daily cost collection across all GCP projects, expensive-query and cache-miss alerts, BigQuery freshness and quota alerts, pipeline summaries, and a central observability dashboard.

Designed a custom MCP server, a governed data-access API for AI-assisted analytics on BigQuery, providing schema discovery, plain-language querying of business metrics, restricted to gold tables via a security layer. Now extending it into active cost governance with dry-run cost gating and a read-only optimization advisor.

Mentor engineers and drive reuse of platform patterns to reduce technical debt.

Apache Iceberg BigQuery Scala / Spark dbt Terraform Pub/Sub MCP

Senior Data Engineer Level 2

Pathao Ltd

Dhaka, Bangladesh Jun 2018 - May 2021

Designed and operated batch and near-real-time telemetry pipelines (trip lifecycle, GPS location streams, driver/rider actions) for a logistics network handling 100K+ daily trips across courier delivery, ride-sharing, and food delivery, spanning multiple cities across Bangladesh and Nepal.

Implemented data pipelines integrating transaction data from 20+ sources plus streaming data at high volume (approximately 50 GB/day) and velocity (peaks of 10K events/sec) on Spark, Dataflow, GCS, and Airflow.

Created the metrics framework computing 100+ operational and financial KPIs (driver earnings, leaderboards, fleet and delivery operations) with under-5-minute end-to-end latency.

Spearheaded the company's data lake on Google Cloud Storage and 10 TB BigQuery data warehouse, halving data retrieval times.

Established the analytics ecosystem on Dataproc, Dataflow, BigQuery, and Data Studio, and delivered a scheduled reporting framework generating 10+ financial and operational reports for business stakeholders.

Dataproc BigQuery Dataflow Python

Senior Data Engineer

Augmedix Bangladesh Ltd.

Dhaka, Bangladesh Jun 2015 - Mar 2018

Core contributor to the design and build of a SQL Server 2016 data warehouse on AWS (S3-backed storage) supporting company-wide business reporting.

Developed AWS data pipelines (S3, Lambda) consolidating data from third-party APIs, Google Sheets, and MySQL (100+ attributes), and a reporting suite producing 20+ operational reports.

Delivered backend services and APIs serving millions of warehouse records to internal web applications, owning quality across the API and data layers.

SQL Server MySQL Data Pipeline APIs

Software Engineer

Nascenia Ltd

Dhaka, Bangladesh Jun 2013 - May 2015

Launched a B2B agricultural marketplace platform with Ruby on Rails and MySQL.

Developed a social media analytics and post-scheduling web application with Rails and PostgreSQL.

Ruby on Rails MySQL PostgreSQL Web Development

Software Engineer

Right Brain Solution Ltd

Dhaka, Bangladesh Nov 2011 - May 2013

Co-authored a high-traffic automotive web portal for the Toronto Star.

Created APIs for social media mobile apps and a multilingual B2C ski-rental e-commerce platform.

API Development Web Applications Mobile Apps OMR Systems