Akshat Singh Chandel · Data Engineer
You're the data.
That little ball is your visit to this page. Run it through the kind of system I build, and see what a data engineer actually does.
Scroll to start
or hold the right arrow key
This is you
engineers call this: a raw eventYour visit, exactly as your browser reported it. Messy, and sent twice by mistake.
Collected
engineers call this: ingestionThe first machine picked you up and stamped where and when you arrived. It changed nothing you sent.
Why: So there is always a record of what really came in.
Keep the original
engineers call this: bronzeInside the tower. Step one: store an untouched copy.
Why: If a later step goes wrong, it can be redone from this.
Clean it up
engineers call this: silverA long web address becomes a site name. A browser code becomes a kind of device.
Why: Reports need plain values, not raw codes.
Fix the time
engineers call this: silverThe time was just text. Now it is a real time that everyone agrees on.
Why: So visits from different time zones can be compared.
Remove the copy
engineers call this: silverYou were sent twice. One copy is kept.
Why: So you are not counted twice.
Add it to the total
engineers call this: goldYou stop being a row and become part of a number.
gold.visits_by_referrerSample count
Why: Totals are what people actually ask for.
On the dashboard
engineers call this: reportingA report reads that total. The green sliver is your visit.
Sample count
Why: Nobody has to build the report by hand.
Answered
engineers call this: an AI agentAn AI agent reads the same total and replies in plain English.
Where did this visitor come from?
read: gold.visits_by_referrer
Why: Anyone can ask, not only people who write code.
One pipeline per source. Never one per table.
The first copy is a receipt. I keep it, I don't clean it.
When two systems disagree, I find the row.
The best report is one nobody rebuilds by hand.
Level clear.
Nuts · Score
One visit, five machines, one answer. That is the job I do.
Your record at every step of the pipeline
This is you
Your visit, exactly as your browser reported it. Messy, and sent twice by mistake.
- Came from:
- Device:
- Arrived:
- Copies: 2 (sent twice, simulated)
Collected
The first machine picked you up and stamped where and when you arrived. It changed nothing you sent.
- Stamp:
- Collected:
- Came from:
- Arrived:
- Copies: 2
Why: So there is always a record of what really came in.
Keep the original
Inside the tower. Step one: store an untouched copy.
- Came from:
- Device:
- Arrived:
- Copies: 2
Why: If a later step goes wrong, it can be redone from this.
Clean it up
A long web address becomes a site name. A browser code becomes a kind of device.
- Came from:
- Device:
- Arrived:
- Copies: 2
Why: Reports need plain values, not raw codes.
- Came from:
Fix the time
The time was just text. Now it is a real time that everyone agrees on.
- Came from:
- Device:
- Arrived:
- Copies: 2
Why: So visits from different time zones can be compared.
Remove the copy
You were sent twice. One copy is kept.
- Came from:
- Device:
- Arrived:
- Copies:
21
Why: So you are not counted twice.
Add it to the total
You stop being a row and become part of a number.
Sample count. Visits from :
127128Why: Totals are what people actually ask for.
On the dashboard
A report reads that total. The green sliver is your visit.
Sample count. Visits from :
127128Why: Nobody has to build the report by hand.
Answered
An AI agent reads the same total and replies in plain English.
Where did this visitor come from?
read: gold.visits_by_referrer
Why: Anyone can ask, not only people who write code.
The 60-second version
Experience
Novigo Solutions Oct 2024 to now
Programmer Analyst (Data Engineer), promoted Jul 2026
- Moved a retail group from Azure Synapse to Databricks on a 10-developer team.
- Helped replace about 900 Data Factory pipelines with about 30 metadata-driven ones.
- Owned one group company's data platform end to end, feeding 6 Power BI reports.
- Built multi-agent and voice-agent demos on Databricks for clients.
Infosys Jun 2022 to Oct 2024
Digital Specialist Engineer
- Built Data Factory and Databricks pipelines for a pharmaceutical client.
- Wrote reusable PySpark templates that teammates used across workstreams.
BE, Information Technology 2022
Thakur College of Engineering and Technology, Mumbai. GPA 9.00 of 10.
What I work in
Certified: Azure Data Engineer Associate (DP-203, 2024) and Azure Fundamentals (AZ-900, 2023).
Reach me
Mumbai, India. Open to remote and hybrid roles.