Sr MongoDB Performance Engineer

Role Descriptions:
Must Have Skills:

1. Expert in MongoDB
2. Strong experience in Performance tuning
3. Fluent in the aggregation framework, indexing strategy, and read/write/read-concern semantics.
4. Strong explain-plan and profiler-driven query optimization
5. WiredTiger internals: cache, eviction, checkpoints, journal, compression, document-level concurrency.
6. Replication and sharding operations at scale, including shard-key design trade-offs.
7. Scripting for automation - Python
8. Strong experience with tuning of specific clusters, queries, indexes and schemas
9. Good experience designing sharding strategy and shard keys; plans resharding and zone strategy
10. Strong experience operating and monitoring existing sharded clusters; runs balancer, fixes hot chunks

Nice To Have Skills:

1. MongoDB certification
2. Kubernetes operator experience for stateful MongoDB; Infrastructure-as-Code (Terraform/Ansible).

Technical/Functional Skills:

MongoDB Performance Engineer to own the throughput, latency, scalability and operational reliability of MongoDB landscape. This is a hands-on database engineering role — not an application-developer role.
The engineer is accountable for making MongoDB fast, predictable and cost-efficient at scale.
Performance engineering (Primary)
• Profile and optimize slow queries and aggregation pipelines using explain plans, the database profiler, and log analysis.
• Design, review and rationalize indexes (compound, partial, wildcard, Time To Live (TTL), text, geospatial) applying the Equality, Sort, Range (ESR) rule; eliminate unused and redundant indexes that inflate write cost and cache pressure.
• Tune the WiredTiger storage engine: internal cache sizing, eviction and checkpoint behavior, compression, and journaling, balanced against filesystem cache.

Operations & reliability
• Operate and scale sharded clusters: balancer management, chunk/range distribution, jumbo-chunk remediation, and (on modern versions) resharding.
• Own backup/restore and Point-In-Time Recovery (PITR) strategy, plus rolling upgrades and patching with zero or minimal downtime.
Infrastructure, OS & platform
• Right-size EC2 instances and EBS volumes (Input/Output Operations Per Second (IOPS) and throughput provisioning); on Kubernetes, own StatefulSet storage class, resource requests/limits, anti-affinity, and cache sizing against container memory limits.

Partner with application teams on data modeling; review schema changes; publish standards, runbooks and capacity guidance.