Nikita FurletovData Platform Engineer
About

Nikita Furletov

Data Platform Engineer · analytics platforms on ClickHouse, S3 and Iceberg, SQLMesh, AWS


y work is building analytics infrastructure on open source — ClickHouse, S3 and Parquet, SQLMesh, Airflow — and I like doing it from an empty repository. I dig into system internals and architectural trade-offs before picking a tool, then own the result: pipelines, models, data quality, catalog, and the infrastructure code underneath.

That works in both directions. At TKB Investment Partners I re-architected a platform nobody trusted into the company's source of truth. On my own time I built a serverless lakehouse for a bookstore chain, four iterations, the latest on AWS. I also run the company's internal AI infrastructure — open-source LLM serving and RAG — because it turned out to be the same discipline: pick carefully, automate everything, keep it observable.

I work as a digital nomad, with Georgia as the first base, on UTC+4.

A platform is trustworthy when nobody checks the number twice.

Where I've worked

2025 — now
Data Platform Engineer, TKB Investment Partners
Company data platform, StockHouse, internal AI infrastructure.
2024 — 2026
Serverless data lakehouse, four iterations
Part-time, continuing the platform I built there as an employee. Finished with the AWS iteration; runs unattended.
2024 — 2025
Data Engineer, Aston
ClickHouse data-mart layer of a DWH; Spark and Airflow ETL for analytics and ML.
2023 — 2024
Data Engineer, bookstore chain
Automated analytics on PostgreSQL and Airflow, surfacing weak points in inventory management.
2022 — 2023
Data Engineer, Institute of Radar Meteorology
Dask/NumPy and Hadoop pipeline plus a PyTorch model, lifting F-score from 0.25 to 0.65.