Benchmark Engineer (AI)

TL;DR

Benchmark Engineer (AI): Designing and maintaining reproducible benchmarks for a vector database with an accent on evaluating performance across different dimensions like latency and throughput. Focus on comparing hirify.global against alternative solutions in a fair and transparent way, and translating benchmark results into clear narratives for product decisions.

Location: Fully remote

Company

hirify.global is an open-source vector database built for high-performance similarity search and AI applications.

What you will do

• Design and maintain reproducible benchmarks for vector search, indexing, filtering, and distributed workloads.
• Evaluate performance across different dimensions: latency, throughput, recall, memory usage, and cost.
• Compare hirify.global against alternative solutions in a fair, transparent, and technically sound way.
• Build and maintain benchmarking tooling, datasets, and automation (CI, dashboards, reports).
• Collaborate closely with core engineers to identify regressions, bottlenecks, and optimization opportunities.
• Help translate benchmark results into clear narratives for docs, blog posts, and talks.

Requirements

• English: B2 required
• Strong software engineering background (Rust, Python, Go, or similar).
• Solid understanding of databases, distributed systems, or search engines.
• Experience with performance testing, profiling, and benchmarking.
• Ability to reason about trade-offs (speed vs accuracy, memory vs latency, etc.).
• Comfort working with large datasets and automation pipelines.
• Clear communication skills to explain numbers and their implications.

Nice to have

• Experience with vector search, ANN algorithms, or ML infrastructure.
• Familiarity with cloud environments and containerized workloads.
• Experience contributing to open-source projects.
• Knowledge of observability tools and performance profiling.

Culture & Benefits

• Work on core infrastructure for modern AI systems.
• Open-source, engineering-driven culture.
• Flexible working hours.
• High ownership, real impact, and technical depth.
• Opportunity to shape how the industry evaluates vector databases.