Mission

• Collect performance metrics (queries, I/O, CPU, memory) across database branches (PostgreSQL, ClickHouse, MySQL, MongoDB).
• Build performance profiles that classify workload types (OLTP, OLAP, mixed) and map them to optimal resource configurations.
• Develop heuristics or ML models that recommend tuning actions (indexing, partitioning, memory settings) based on the profile.
• Integrate profile generation and tuning recommendations into the agent workflow: branch → profile analysis → apply tuning.
• Collaborate with engineering teams to embed performance profiles into Guepard’s branching and time-travel infrastructure.
• Document profile definitions, tuning strategies, and profiling results for visibility and audit.

Primary responsibilities

• Design and implement data collection pipelines to gather query-level and system metrics from multiple DB engines (Postgres, ClickHouse, MySQL, MongoDB).
• Define feature representations for workloads and implement profiling logic to classify workload types and resource needs.
• Research and implement ML models or rule-based heuristics to predict optimal tuning actions (index suggestions, partition strategies, memory/config tunings).
• Integrate the profile generation and recommendation components into the existing Performance Optimization Agent workflow.
• Write clear documentation of profile schemas, tuning rules/models, and experimental results for audit and operational use.

Deliverables & collaboration

• Deliver a working Performance Profile component that can (a) ingest metrics from branches, (b) produce workload profiles, and (c) output tuning recommendations.
• Provide evaluation reports comparing baseline vs. recommended configurations (latency, throughput, resource utilization) and include test cases.
• Collaborate closely with backend and infra engineers to ensure safe application of tuning actions within branching/time-travel infrastructure and CI/CD constraints.
• Implement monitoring/validation to detect regressions and provide rollback or safety checks when applying automated tunings.

How to apply

• To apply, send your application to jobs@guepard.run with the subject line indicated below.
• You can also apply via the online link: https://lnkd.in/d8wuxwT3