Database Operations
Own the data layer across our client platforms. You'll design, deploy, and operate database infrastructure spanning ChromaDB for vector workloads and ScyllaDB for high-throughput transactional systems.
About the Role
We're looking for a high-agency operator who thinks in data models, lives in terminal sessions, and has strong opinions about consistency vs. availability trade-offs. This role sits at the intersection of infrastructure and application engineering; you'll design schemas, optimize queries, tune clusters, and keep our data layer fast and reliable across multiple client deployments. It's 2026, and we expect you to operate like it: drive Codex and Claude Code to write migrations, tooling, load-test harnesses, and runbooks far faster than by hand, then verify rigorously, because in data infrastructure an unverified change is a future incident. The leverage is huge and so is the blast radius, so the engineers who thrive here pair fast AI-assisted iteration with disciplined testing: schema changes rehearsed against realistic data, migrations dry-run and reversible, performance claims backed by benchmarks rather than vibes. ChromaDB and ScyllaDB are core to how we build. ChromaDB powers our vector search and embedding pipelines for AI-augmented features; ScyllaDB handles the high-throughput, low-latency workloads our fintech and food-tech clients depend on (millions of writes per second at single-digit-millisecond p99). You'll own both, and you'll own the verification that keeps them trustworthy.
What You'll Do
Design and maintain ScyllaDB clusters and ChromaDB instances for high-throughput transactional and vector workloads across client platforms
Use Codex, Claude Code, and similar tooling to move fast on migrations, automation, load-test harnesses, and runbooks, then review and verify every change before it touches production
Build the testing and verification infrastructure for the data layer: realistic test datasets, migration dry-runs, reversible changes, and reproducible benchmarks
Define data models, partition strategies, and compaction policies for ScyllaDB; build collection schemas, indexing, and query optimization for ChromaDB
Back every performance and capacity claim with benchmarks and load tests rather than assumptions
Set up monitoring, alerting, and capacity planning, and treat code review as a primary gate, especially for AI-authored migrations and tooling
Write runbooks, conduct blameless incident post-mortems, and continuously raise operational reliability
Evaluate and benchmark new database and AI tooling as workload requirements evolve
Requirements
3+ years operating distributed databases in production (ScyllaDB, Cassandra, or DynamoDB)
High agency: you own the data layer end to end, unblock yourself, and drive reliability without being asked
Daily, fluent use of AI coding tools (Codex, Claude Code, or similar) to move fast, paired with the discipline to verify and review what they produce
A rigorous testing and verification habit applied to data: rehearsed migrations, realistic test data, reversible changes, and benchmark-backed claims
Hands-on experience with vector databases (ChromaDB preferred; Pinecone/Weaviate/Milvus acceptable)
Strong understanding of the CQL data model, partition design, and ScyllaDB internals (compaction, repair, streaming)
Comfortable in Linux, shell scripting, infrastructure-as-code (Terraform/Ansible/Pulumi), and container orchestration (Docker, Kubernetes), with solid distributed-systems fundamentals (CAP, eventual consistency, consensus)
Nice to Have
You build your own automation, evals, or agents to operate and verify infrastructure faster
Experience with ScyllaDB Alternator (DynamoDB-compatible API)
Contributions to open-source database projects
Familiarity with LangChain, LlamaIndex, or similar frameworks that integrate with ChromaDB
Performance tuning experience at scale (>1M ops/sec) backed by reproducible benchmarks