Job Title: Software Engineer, Data Pipelines (Mid-Level)
Location: Remote, Ukraine
Team: Data & Platform Engineering — working side-by-side with our Data Science team
Overview
SeeTree is a leading company in the Ag-tech industry, providing per tree intelligence platform to growers to track their trees’ health and productivity.
Using rich sources of information such as drones, satellites, IOT sensors, weather information and more, we can scan and analyze hundreds of millions of trees and provide deep information on each and every one of them!
Our mission is to boost growers and industry ROI by digitally transforming agronomy, operation and decision making.
🎯 The Role
We are looking for a mid-level engineer to own the backend data pipelines that power our Data Science team. You will sit between Data Science and our GCP data platform: building the pipelines that produce the data our models learn from, and taking the models and prototypes the DS team builds and making them run reliably, on schedule, at continental scale.
This is a hands-on Python role. Concretely: satellite and drone imagery lands in cloud storage, and a chain of Airflow DAGs turns it into per-tree and per-field facts in BigQuery — ingestion, raster processing, alignment, detection, feature extraction, model inference, publication. You will build, extend and operate that machinery, and make it a platform the data scientists can move fast on instead of a black box they file tickets against.
🛠 What You’ll Do
Build and maintain batch data pipelines in Python, orchestrated with Apache Airflow on Cloud Composer 2 — DAGs, custom operators and sensors, retries, backfills, SLAs
Productionize data science work: take notebooks, prototypes and trained models from the DS team and turn them into containerized, scheduled, monitored jobs on GKE and Cloud Run
Design and evolve BigQuery datasets — feature tables, training sets, partitioning and clustering strategies, and cost / performance tuning of heavy analytical SQL
Own pipeline reliability: idempotent writes, checkpointing and resume for long historical backfills, data-quality checks, alerting, and failure modes that are loud rather than silent
Work with large geospatial datasets — Sentinel-2 imagery, drone orthomosaics, elevation models — computing per-field and per-tree statistics that feed downstream models
Build and maintain the Python services around the pipelines (Cloud Run, Flask / FastAPI) that expose results to our applications, dashboards and to the data scientists themselves
Contribute to architecture, specs and code reviews across a multi-repo, service-oriented GCP platform
⚙️ Our Stack
You do not need every line of this on day one — but this is what the work actually looks like:
Languages: Python (primary), SQL, some TypeScript / Node.js
Orchestration: Apache Airflow on Google Cloud Composer 2
Data: BigQuery, Google Cloud Storage, Firestore, Pub/Sub
Compute: Kubernetes (GKE) Jobs, Cloud Run, Cloud Functions, Docker
Python libraries: pandas, NumPy, GeoPandas, Shapely, rasterio, PyTorch (models we run)
Infra & CI: Terraform, Bitbucket Pipelines, Cloud Monitoring
AI tooling: Claude Code, Cursor, MCP servers, spec-driven agent workflows
✅ What We’re Looking For
3+ years of professional experience writing production Python (data-intensive backends, ETL / ELT, or platform work)
Hands-on experience with a workflow orchestrator — Airflow strongly preferred; Dagster, Prefect or equivalent is fine if you are willing to learn Airflow properly
Strong SQL and real experience with a cloud data warehouse — BigQuery ideally, or Snowflake / Redshift / Databricks. You can look at a query and explain why it costs what it costs
Comfortable on a major cloud (GCP preferred; AWS / Azure transferable) and with containers — you can write a Dockerfile and debug a failing Kubernetes job
Experience collaborating with data scientists or ML engineers — you have taken someone else’s model or analysis
