Data Scientist (Maritime Analytics & Remote Sensing)
Description
About the Role
We are looking for a Data Engineer to design, build and run the data platform behind our maritime products: live vessel tracking, marine traffic and risk studies, and the data science and AI systems built on them. You will own the whole path, from the edge server at a receiver site, through stream processing, to the analytical and operational databases, together with the servers and networks it runs on. The role is hands-on, and part of it is on site: installing servers, connecting them to the network and bringing them up from a bare operating system.
Key Responsibilities
Streaming Ingestion & Pipelines
- Build and run publishers that move raw AIS/NMEA streams from edge servers and external feeds into the streaming platform, with local buffering, reconnect with back-off, and no silent message loss.
- Administer the streaming platform: topic, partition and key design, retention, consumer groups and offsets, authentication and access control.
- Develop and operate stream-processing jobs for decoding, enrichment, latest-position state and event detection (anchoring, berthing, zone entry and exit, proximity), handling event time, late data, checkpoints and backpressure.
- Build incremental, idempotent batch pipelines, and run backfills and reprocessing from the raw archive.
- Measure freshness, lag and loss at every stage, and trace a missing record to the stage that dropped it.
Data Architecture & Geospatial Modeling
- Design raw → staging → curated models for trajectories, reference data (ports, anchorages, zones, boundaries) and spatial outputs (geofencing, proximity, grid aggregates), covering geometry storage, spatial indexing, partitioning, clustering and naming conventions.
- Plan and execute large-scale data migrations with change data capture, reconciliation and minimal-downtime cutover.
Data Warehouse & Databases
- Write advanced SQL to pull data for analysts, data scientists and client studies, and turn repeated pulls into curated tables or parameterized templates.
- Manage warehouse datasets (partitioning, clustering, expiration, schema evolution, access control, data catalog), and keep loads complete, deduplicated and reconciled with source counts.
- Control cost and performance: query cost estimates and limits, materialized and incremental tables, and spend monitoring in the warehouse; indexing, query plans, maintenance, replicas and retention in operational databases.
Infrastructure & Networking
- Set up servers end to end at client premises, data centers and receiver stations: rack and cable them, connect them to switches, configure IP addressing, VLANs, routing, DNS, NTP and firewalls, and install and harden the operating system.
- Deploy and support on-premise installations, working within each client’s security and change-control rules.
- Run cloud infrastructure (virtual machines, managed databases, object storage, DNS, CDN and certificates) and containerized services that come back by themselves after a reboot; plan capacity and test backups and restores.
Reliability, Data Quality & Governance
- Lead CI/CD, infrastructure automation, environment separation and version-controlled schema migrations.
- Run monitoring, alerting, SLAs and runbooks; lead incident response and write post-incident reviews that prevent repeats.
- Own the data quality framework (automated checks, anomaly detection, reconciliation, quality SLAs), and benchmark data sources against each other on coverage, latency and completeness.
- Enforce governance: ownership, lineage, access control, retention and secrets management across all data stores.
AI Enablement & Collaboration
- Build curated datasets, knowledge graphs and data interfaces that let agentic AI tools query maritime and geospatial data reliably, with guardrails on quality, lineage and access.
- Partner with data scientists on study data and production models, mentor others, and document architecture, tables and runbooks.
Qualifications
Must-Have
- Degree in Computer Science, Software or Data Engineering, or equivalent experience.
- 5+ years building and operating production data pipelines, including streaming.
- Fluent in Python, plus at least one of Java, Scala, Rust or Go.
- Advanced SQL with hands-on BigQuery and PostgreSQL: schema design, partitioning, indexing, query tuning and cost control.
- Kafka in production, and a stream-processing framework (Flink preferred, in Java or PyFlink; Kafka Streams, Spark Structured Streaming or Beam also considered).
- Workflow orchestration with Airflow, cron or similar schedulers.
- Linux administration and networking (TCP/IP, subnets and VLANs, routing, DNS, firewalls, VPN, TLS); able to take a server from bare metal to a running service.
- Sound engineering practice: idempotency, replayability, observability, Git, CI/CD, Docker, and data quality checks in production pipelines.
- Experience with geospatial data and analytics-ready data models.
- Willing to travel to client sites and receiver stations; government projects may require security vetting.
Preferred
- Streaming and storage: Flink state, checkpoint and backpressure tuning; PostGIS, TimescaleDB, H3 and spatial indexing (GiST, BRIN).
- Domain: AIS/NMEA pipelines, or similar telemetry streams from edge or IoT devices.
- On premise: managed switches, RAID, out-of-band management (IPMI, iDRAC, iLO), and government or regulated networks.
- Platform: Terraform or Ansible, Kubernetes, Prometheus/Grafana, GCP (Cloud Run, Dataflow, Composer), CDC (Debezium, Kafka Connect), Keycloak/OIDC.
- AI: knowledge graphs and ontology design; AI-assisted development (e.g. Claude, Codex) in a production workflow.
What We Look For
- Owners, who treat an outage or a data gap as theirs to find, fix and prevent.
- Comfortable moving between a server rack, a Kafka topic and a SQL query.
- Architecture-first: clear conventions, and a sound balance of cost, performance and maintainability.
- Growth-minded and proactive in raising engineering standards in a fast-moving team.
Skills
- Agentic AI
- AI
- Ai Enablement
- Airflow
- Analytics
- Anomaly Detection
- Ansible
- Authentication
- Automation
- BigQuery
- CI/CD
- Cloud
- Cloud Run
- Data Engineering
- Data Pipelines
- Data Quality
- Data Science
- Data Warehousing
- Debezium
- DNS
- Docker
- Firewall
- Flink
- GCP
- Git
- Grafana
- IoT
- Java
- Kafka
- Keycloak
- Kubernetes
- Linux
- Networking
- Observability
- OpenID
- PostgreSQL
- Prometheus
- Python
- Rust
- Scala
- Secrets Management
- Spark
- SQL
- TCP/IP
- Terraform
- TimescaleDB
- TLS
- VLAN
- VPN
- Workflow Orchestration
Sign in to read and join the discussion.