Staff Data Platform Engineer

Tatari•San Francisco, CA

10h•Hybrid

About The Position

This is a systems and infrastructure position first. As a Data Platform Engineer, you will be responsible for the reliability, stability, and operational health of our data platform — including how it is deployed, monitored, maintained, and promoted across environments. Data engineering skills are a plus and will be developed on the job; what we cannot teach is operational discipline. If you have spent your career keeping production systems alive, know what it feels like to break prod and never want to do it again, and treat lower environments as non-negotiable gates rather than suggestions — we want to talk to you. This is not a data engineering role. You will not spend most of your time writing jobs or consuming the platform. You will be administering, scaling, hardening, and evolving it.

Requirements

Operational instinct — "the fear" — you've been burned by prod, you respect it, and you've built habits around it. You know what a proper maintenance window looks like, you communicate before you touch production, and you don't spin up new initiatives while something critical is still burning in.
3+ years in cloud infrastructure, SRE, or platform engineering (AWS preferred; GCP/Azure experience translates)
High Availability architecture: blue/green deployments, data replication, load balancing
Experience with workflow orchestration (Airflow or similar DAG-based schedulers — or general job scheduling/cron systems at scale)
Strong Linux fundamentals and scripting (Bash, Python, or similar)
Distributed data processing (Spark, PySpark, or similar big data frameworks — or experience managing clusters that run them)
Containerization and orchestration (Kubernetes, Docker, or similar)
Data ingestion, ETL, or streaming systems (Kafka, Flink, or similar — or experience operating message queues and pipelines)
Infrastructure-as-code and provisioning (Terraform, Helm, or similar)
OLAP and OLTP databases (Clickhouse, Postgres, Redshift, or similar — query patterns, indexing, and operational care)
Monitoring, logging, and observability (Datadog, Prometheus, Kibana, or similar)
Managed data platforms (Databricks or similar — administering and scaling, not just consuming)
Network infrastructure fundamentals : load balancers, DNS, auto-scaling, multi-region topologies, proxies
Security and access management : least-privilege, secrets management, controls for data systems
Humility — you don't know everything, you say so, and you ask before acting in unfamiliar territory
Methodical execution — you minimize variables, you don't premature-optimize, you finish what you started before starting something new
Communication — you tell the team what you're doing before you do it, especially in shared or production environments
Ownership — when something goes wrong, you look inward first
Independence – you can drive projects end-to-end, from ambiguous requirements to high quality deliverables. But you aren’t afraid to ask for help.

Nice To Haves

MLOps concepts or tooling

Responsibilities

Own the reliability and availability of our data platform infrastructure across all environments
Enforce and improve environment promotion discipline — staging is not prod, and prod is sacred
Define and uphold SOPs around deployments, maintenance windows, and change management
Instrument and monitor platform health using observability tooling; build alerting that means something
Participate in architecture and deployment discussions; push back when something isn't ready
Collaborate with data scientists, engineers, and product managers on infrastructure needs — as a partner, not an order-taker
Identify and remediate reliability risks before they become incidents
Support customer-facing and internal systems with a bias toward stability over velocity