Ishtiak Alam
Data and Analytics Engineer with 12+ years in fintech, SaaS, telecom and healthcare. I work on Databricks, dbt, Snowflake and Azure, from the first ingestion job to the dashboard your finance team signs off on.
Taking on select consulting projects
Data engineers make data arrive. Analytics engineers make it mean one thing. The handoff between the two is where most numbers break, so I do both and the handoff disappears.
Data engineering
Analytics engineering
Results from production systems at a regulated FX broker and a global SaaS company, with full write-ups for three of them. Client data stays confidential; the numbers are real.
Showing all 6 projects
Moved 25+ finance and FP&A reports to Databricks in six months, with automated reconciliation proving every number matched before anything was switched off.
Read the case studyHow it’s measured: yearly SQL Server licenses and servers, plus the manual cross-checks, retired along with the old platform.
Rebuilt payments ingestion from a daily full reload into an incremental hourly job, so finance and risk stopped working from yesterday’s numbers.
Read the case studyJob run time
How it’s measured: data delay is the longest a payment can wait before it shows in reporting; run time is the job’s wall-clock duration.
A streaming consumer that lands every trade and account event in the lakehouse, feeding intraday exposure and fraud dashboards.
How it’s measured: time from an event arriving on Event Hubs to being queryable in Delta Lake.
Risk and fraud teams needed to see exposure during the trading day, not after the nightly batch.
Built a Spark Structured Streaming consumer on Azure Event Hubs through its Kafka API, writing to Delta Lake so exposure and fraud dashboards read events within two minutes.
Spark Structured Streaming, Azure Event Hubs, Kafka API, Delta Lake
Sales, marketing, finance and compliance now read the same 45+ metrics from one dbt MetricFlow semantic layer, used by 120+ people.
Read the case studyHow it’s measured: overlapping reports and spreadsheets retired once their numbers came from the semantic layer.
Clients hold several trading accounts, so acquisition cost was being counted per account. Measuring per person changed where the budget went.
How it’s measured: paid media spend divided by clients whose first deposit cleared, counted per person, before and after the budget shift.
MT4 and MT5 accounts belonging to the same person were counted separately in acquisition reporting.
Resolved every account to one Salesforce client ID, defined New Traded Client, first funding and client P&L at client grain with Marketing, and built cost per acquisition by channel. Three underperforming campaigns were cut.
dbt, Databricks, Salesforce data, Power BI
Architected the central Snowflake warehouse that became the single source of truth for 40+ SaaS KPIs, used by 80+ people in finance, product, growth and sales.
Access and PII incidents a quarter
How it’s measured: access and PII incidents per quarter, before and after the Atlan rollout.
Five disconnected source systems gave five different answers for ARR, churn and active users.
Built the warehouse with Fivetran ingestion and Airflow orchestration, consolidated the sources into one Kimball model, and rolled out Atlan for cataloging, PII tagging and role-based access. Power BI semantic models removed 40% of recurring report requests.
Snowflake, Fivetran, Airflow, Python, Power BI, Atlan
I take on a small number of consulting projects at a time. Each one starts with a free 30-minute call to check that I’m the right person for the problem.
Fixed fee, 2 weeks
A review of your Databricks or Snowflake setup: cost, reliability, governance and how the team ships changes.
You get a written report with findings ranked by impact, and a 90-day plan your team can run without me.
Fixed scope, 4 to 8 weeks
Metric definitions agreed with the people who own them, dbt models with tests and CI, and a semantic layer your BI tools read from.
You get one set of numbers every team uses, with the code, docs and training to keep it that way.
Project, scoped per estate
Moving reporting off SQL Server, Oracle or Teradata to Databricks or Snowflake, with automated reconciliation and business sign-off before anything is switched off.
You get a retired legacy platform, and a finance team that trusts the new one from day one.
Monthly retainer
Architecture decisions, code review, hiring help and a data roadmap for teams that need senior direction but not a full-time lead yet.
You get standards, reviews and a roadmap, with a set number of hours each month.
60 minutes, paid
A working session on one hard problem: a slow pipeline, a disputed metric, a design you want a second opinion on.
You get a direct answer and written notes after the call.
30 minutes, free. You describe the problem; I tell you honestly whether I can help.
Scope, deliverables, timeline and a fixed price, in writing.
Work happens in your code and your cloud, with a demo every week.
Documentation, runbooks and a training session, so your team owns it.
I’m based in Dhaka (UTC+6), which overlaps the working day in Singapore, Australia and the Gulf, and the morning in Europe. Everything happens remotely.
I’m used to regulated environments: ASIC, EMIR, MiFID, MAS and CySEC reporting, and PHI under HIPAA. NDAs are no problem.
Free, no preparation needed. Bring the problem, and you’ll leave with a clear next step whether or not we work together.
Md Ishtiak Alam, Dhaka, Bangladesh
I’m Ishtiak, a data engineering lead based in Dhaka, Bangladesh. For 12 years I’ve built the systems companies use to count money, customers and risk.
I started as a data warehouse and BI developer at IQVIA, building ETL and dashboards for pharma clients. At Banglalink, a telecom operator with 32 million subscribers, I ran the BI team and caught fraud that was costing about USD 190K a month. At Optimizely I architected the Snowflake warehouse that gave the company one answer for ARR and churn.
Today I Data and Analytics Engineering at Axi, a global FX and CFD broker regulated in five jurisdictions. I own the trading, payments, customer, marketing and finance domains on Databricks, dbt and Azure: 45 production jobs, 80+ dbt models, a semantic layer used by 120+ people, and the regulatory reporting that has to be right every day.
I’ve led teams of 4 to 10, mentored 3 engineers and 15+ analysts, and written the standards other people build on. The work I like best is getting a room to agree on what a number means, then making sure it never drifts.
2023 – now
Data and analytics engineering for trading, payments, customer, marketing and finance. Remote.
2022 – 2023
Central Snowflake warehouse, dimensional model and data governance.
2022
Pharma supply-chain analytics; automated 90% of manual ETL steps.
2018 – 2022
Led a 4-person BI team serving 150+ users across every department.
2014 – 2018
Data warehouse, ETL and Tableau dashboards for pharma clients.
Code, dashboards, writing and career history live in a few places. Here’s what you’ll find in each.
Open-source reference build on the stack I run in production
The repository goes public soon. Follow me on GitHub to see it first.
Consulting projects, expert sessions, digital products, or a question about your data stack. I reply within two working days.