AI Doesn't Understand Your Codebase (And That's Your Fault)
How we made AI coding assistants actually useful on a 500-model dbt project by documenting everything.
AIdbtDocumentation+1
Jan 15, 20264 min read
I'm Ashyam, a problem-solver
who makes data trustworthy.
7+ years building analytics platforms and data infrastructure for companies that actually care about their data (and a few that learned to care the hard way).
Currently Lead Data Engineer at Saras Analytics, leading a team of 4. We build data products for e-commerce clients who need their numbers to mean something.
I specialize in the modern data stack — Python, dbt, BigQuery, Airflow — and occasionally wander into React when nobody else will fix the dashboard.
Location-flexible. Open to remote, hybrid, or wherever the coffee is good.
PythonSQLdbtBigQueryAirflowGCPDockerReact
Where I've Shipped Real Code
7+ years of turning “it works on my machine” into “it works in production.” Most of the time.
Jan 2025 - Present
Lead Data Engineer
Saras AnalyticsLeading the data engineering function for Pulse. Driving architecture, automation, and AI-assisted engineering tools to make data infrastructure self-operating.
Lead Data Engineer
Jan 2025 - Present
Saras AnalyticsLeading the data engineering function for Pulse. Driving architecture, automation, and AI-assisted engineering tools to make data infrastructure self-operating.
Aug 2023 - Jan 2025
Senior Data Engineer
Saras AnalyticsDesigned the EDM Insights data model, unifying 20+ platforms into consistent dimensional models. Reduced BigQuery costs by 60%.
Senior Data Engineer
Aug 2023 - Jan 2025
Saras AnalyticsDesigned the EDM Insights data model, unifying 20+ platforms into consistent dimensional models. Reduced BigQuery costs by 60%.
Sep 2021 - Aug 2023
Data Science Engineer
Peninsular Research OperationsBuilt backend data and ML infrastructure supporting analytics for state government projects. Automated model lifecycle with drift detection.
Data Science Engineer
Sep 2021 - Aug 2023
Peninsular Research OperationsBuilt backend data and ML infrastructure supporting analytics for state government projects. Automated model lifecycle with drift detection.
Jan 2018 - Sep 2021
Software Engineer
Algonox TechnologiesContainerized internal tools and automation scripts. Optimized Linux backend workflows with Shell scripting and concurrent Python.
Software Engineer
Jan 2018 - Sep 2021
Algonox TechnologiesContainerized internal tools and automation scripts. Optimized Linux backend workflows with Shell scripting and concurrent Python.
Things I've Built...that actually (kinda) work
Side projects, internal tools, and the occasional “I'll just learn a new framework this weekend” adventure.
Clarity
Personal Finance Platform
PythondbtBigQueryLLM
End-to-end data platform using Python, Airflow, dbt, BigQuery. Features LLM-based transaction categorization and an AI financial advisor for spending insights.
DE Agent
AI Debugging Assistant
Claude SDKCloud RunPython
Built with Claude Agent SDK, deployed on Cloud Run with Slack integration. Reduced RCA time from 1-2 hours to 10-30 minutes across the team.
Pulse Admin
Observability Platform
ReactFirebaseBigQuery
Full-stack tool with React & Firebase monitoring BigQuery costs, DAG health, and data quality. Self-taught React to fill a gap nobody else would.
The Credentials...nobody asks about
Proof that I can sit through lectures and pass exams. The real learning happened after graduation.
2021 – 2022
Master of Science in Data Science
Liverpool John Moores University
Master of Science in Data Science
2021 – 2022
Liverpool John Moores University
2019 – 2020
PG Diploma in Data Science
International Institute of Information Technology
PG Diploma in Data Science
2019 – 2020
International Institute of Information Technology
2013 – 2017
Bachelor of Engineering in Computer Science
Birla Institute of Technology and Science
Bachelor of Engineering in Computer Science
2013 – 2017
Birla Institute of Technology and Science
Thoughts...occasionally worth reading
Let's Build Something Together
Whether you need a data platform, want to chat about pipelines, or just want to argue about Python vs SQL — I'm here.