I used to be the person who had to explain why the dashboard was wrong.
Six years inside fintech and banking systems showed me what bad data costs. Now I build the pipelines that fix it: production ETL, star schema warehouses, and CI/CD-ready infrastructure.
Based in Suffolk, England · Remote or hybrid · UK-based roles
Source FileExtractStageTransformWarehouseBI Layer
Core stackPythonPySparkSQLAirflowPostgreSQLDockerAWSPower BI
A containerised on-premise data platform built for a multinational retail scenario. Five Docker services (PostgreSQL, Spark, Airflow init/webserver/scheduler) brought up by one command. A PySpark medallion pipeline (Bronze partitioned Parquet to Silver joined frame to a three-table dimensional Gold layer) runs end-to-end in 3 minutes 11 seconds at stock scale. Airflow's parameterised DAG supports both scheduled stock runs (analytical truth) and on-demand scaled tests (architecture validation), demonstrated processing 7.5M rows through the pipeline at 2x scale.
A distributed ETL pipeline that processes 1 million synthetic banking transactions through five disciplined stages (extract, explore, clean, transform, load) into a validated PostgreSQL star schema. Deterministic SHA-256 surrogate keys, type-aware cleaning that flows DECIMAL through to the warehouse, validation at every transform stage, and idempotent re-runs that never duplicate a row.
XTD Research Labs: Async Ingestion & PySpark Medallion Pipeline
UK Carbon Intensity API · aiohttp async · PySpark · PostgreSQL
A three-stage pipeline built for a UK grid decarbonization research scenario. Async aiohttp pulls 1,095 days of regional carbon intensity from a live government API into a Bronze data lake, with semaphore-bounded concurrency and exponential-backoff retry. PySpark explodes deeply nested JSON and pivots fuel types across Silver and Gold layers. A two-stage dedup-merge lands the result in PostgreSQL with composite-key idempotency. 8.7M intermediate rows down to 19,728 daily research metrics.
Three more case studies covering different architectural patterns: an end-to-end ETL warehouse on Brazilian e-commerce data, a layered three-schema analytics platform, and a live web scraping pipeline.
PayFlow: ETL & Warehouse
9 source CSVs · star schema · idempotent loads
A production-style ETL pipeline that ingests 9 raw Brazilian e-commerce CSVs, validates and cleans through a multi-layer pipeline, and loads a full star schema into PostgreSQL. Idempotent loads, structured logging, and a single-command orchestrator.
An end-to-end ETL pipeline that turns a flat chocolate sales dataset into a structured PostgreSQL warehouse with raw, operational, and analytics schemas. Includes 10+ engineered features, a full dimensional model, FK validation, and a one-command orchestrator.
Live web scraping · Selenium · feature engineering
A production-style web scraping pipeline that scrapes 60 pages of AliExpress laptop listings, cleans and enriches the data with discount metrics and price bands, and appends only new records to PostgreSQL. Append-only loads, in-memory df passing, and a health check before scraping.
60 pages per run6 engineered featuresAppend-only loads
Analytics engineering project with dbt + BigQuery on public transit data. Materialised models, tests at every layer, generated lineage docs, and a live Looker Studio dashboard. Targeting Q2 2026.
dbtBigQueryLooker Studio
03
Experience
Work history
6+ years across data engineering, fintech product operations, and analytics, with quantified impact from production systems.
Jan 2026 – Present
Amdari Inc.
UK · Remote
Data Engineer Current
Designed end-to-end pipelines using Python, SQL & Apache Airflow. Improved batch processing efficiency by 35%.
Optimised PostgreSQL via partitioning & query tuning. Achieved £10K quarterly cost reduction in AWS S3 and 40% query performance gain.
Automated workflows with Airflow DAGs, eliminating 20+ hours of manual processing per month.
Integrated CI/CD with Docker, Jenkins & GitHub Actions. Reduced release rollbacks by 50%.
Monitored cloud infrastructure with Grafana, maintaining 99.9% uptime.
Built Power BI dashboards cutting executive decision time by 30%.
Built batch data workflows in Python: ingested CSV exports from the e-commerce platform, applied cleaning, validation and standardisation, and produced analysis-ready datasets for reporting.
Produced analytics reports on daily transactions and stock levels, surfacing inventory trends that informed restocking decisions and reduced over-stocking / under-stocking.
Analysed sales data to generate actionable insights on product performance, sales trends, and customer buying behaviour, supporting marketing strategy and operational planning.
Investigated customer enquiries and complaints to detect recurring patterns, recommending process improvements to enhance service quality and retention.
Every project on this site follows a small set of positions I have taken often enough that they feel less like opinions and more like defaults.
Seven principles I bring to every project.
These are not borrowed from blog posts. They are positions I have defended in pull requests, learned from incidents, and applied in production. Each one comes with the trade-off I accept by holding it.
I use TaskGroups for clarity over SubDAGs. SubDAGs share a scheduler slot and quietly cause backfill pain.
On PostgreSQL
My default for analytics warehouses up to ~100M rows. Past that I'd reach for Snowflake or BigQuery before adding complexity.
On Python
I treat type hints as documentation, not safety nets. They make the next engineer's job easier. The runtime doesn't care.
06
Certifications
Credentials
10Analytics
Data Engineering
Jan 2026
Udemy
Diploma in Web Development
Nov 2024
Product School
Product Analytics Micro-Certification
Jan 2023
Alison
Introduction to Tableau Desktop
Jul 2023
Google
Fundamentals of Digital Marketing
Jul 2020
FITC
Basic Banking Operations
Apr 2018
07
Tailored CV
Email me for a tailored CV
Yomi Ismail: Data Engineer
Every CV I send is tailored to the role you're hiring for, mapping my work directly to the stack and responsibilities you need. Send me a quick email with the role title (or a link to the job description) and I'll send back a matched version within 24 hours.
Currently looking for Data Engineer or Analytics Engineer roles. Also open to senior data platform or technically-focused operations positions with a strong data component. Available remote or hybrid across the UK. If you're hiring or have a project you'd like to discuss, feel free to get in touch.