Software Engineer, Data Platform The data backbone behind AI dealmaking. Pipelines, warehouse, tooling. Engineering 4–6+ yrs Python · dbt · Airflow Postgres · AWS View role → →
A data platform engineer builds the ingestion pipelines, warehouse (dbt models, internal APIs) and internal tooling that deliver deal-signal data for VC clients, including entity resolution across sources like PitchBook and Affinity. Core stack: Python, SQL/Postgres, dbt, Airflow/Dagster, and cloud infra on AWS.
- Compensation
- Not disclosed
- Location
- Buenos Aires, Argentina
- Work setup
- Remote
- Seniority
- Not specified
- Employment
- Full time
- Experience
- 6+ years
- Category
- Data engineering
- Posted
Skills
Must have
- 4–6+ years backend or data engineering experience
- Python
- SQL / Postgres
- dbt
- Airflow / Dagster
- Cloud architecture & infrastructure (AWS, Azure or GCP)
- Docker
- Graph databases
- Semantic search
- Claude Code / Codex (AI coding tools)
- Based in Argentina; if near Buenos Aires or Córdoba, office two days a month
Nice to have
- DuckDB
- Kafka
- LLM tooling
Full description
Engineering Remote · Argentina Paid in USD Full-time
Software Engineer, Data Platform
Top-50 VC firms decide where billions go. The signal under those decisions runs on pipelines someone has to build well. That's this job.
The ingestion + warehouse layer, end to end
Experience
4–6+ years backend or data engineering
Compensation
USD, discussed on call 1
Process
Python SQL / Postgres dbt Airflow / Dagster Cloud architecture & infrastructure (AWS, Azure or GCP) Docker Graph databases Semantic search Claude Code / Codex Nice: DuckDB, Kafka, LLM tooling
Fit check
Tick what's true. We'll tell you straight. Nothing is sent anywhere. This is for you, not us. Six statements, honest answer only.
Ingestion jobs, dbt models, internal APIs and the tooling that lets four other engineers move without asking you first.
PitchBook, Harmonic, Affinity, ten more. Reconciling entities that disagree about their own names and funding rounds.
With users
Calls with the people using what you built: analysts, ops leads, partners. You hear the complaint first-hand, not through a ticket.
Review
Reading other people's code and having yours read. No standup theatre.
T h r e e t h i n g s y o u ' d s h i p i n y o u r f i r s t y e a r
01 A signal pipeline a fund runs on daily
Dozens of sources into one warehouse, deduped and current. Sourcing team opens it every morning.
02 Entity resolution that stops the arguing
One company, five spellings, three funding numbers. You decide which one is true, and prove it.
03 Tooling that makes the next build a week, not a month
Shared contracts, tests, local dev that works. The multiplier on everyone else's output.
- Ambiguity is where you start, not where you stop.
- You care that the number is right, even when nobody would notice.
- You want the client relationship, not protection from it.
Be honest, probably not for you if
- You want a spec handed to you and a quarter to deliver it.
- Your favourite part of the job is the architecture diagram.
- Talking to the people who use your work sounds like a distraction.
- You need a large team's process around you to feel safe shipping.
W h a t y o u g e t
Same for every role, every level.
Remote-first
Work from anywhere in Argentina. If you live near Buenos Aires or Córdoba, you're in the office two days a month.