Gimlet Labs is seeking a Member of Technical Staff focused on ML systems and inference. In this role, you will design and build the inference systems that execute full models end-to-end under real production constraints. You will work at the intersection of model architecture, runtime behavior, and system performance to ensure inference is fast, predictable, and scalable.
This role is ideal for engineers who deeply understand how modern models execute in practice and who care about latency, throughput, and memory behavior across the full inference lifecycle.
Responsibilities- Design and optimize end-to-end inference pipelines from request ingestion through execution and response
- Build and evolve inference runtimes that balance latency, throughput, and concurrency under real-world load
- Reason about batching, queuing, and scheduling tradeoffs, including their impact on tail latency and fairness
- Manage KV cache allocation, placement, reuse, and eviction across models and requests
- Optimize prefill and decode paths, including attention mechanisms and memory usage
- Profile and debug inference performance issues across model, runtime, and system boundaries
- Work closely with compilers, kernels, networking, and distributed systems to deliver end-to-end performance improvements
Qualifications- Strong software engineering fundamentals
- Experience building or operating ML inference or model serving systems
- Comfort reasoning about performance, memory usage, and system behavior under load
Preferred Qualifications- Experience with inference runtimes such as TensorRT-LLM, vLLM, or custom serving systems
- Deep understanding of modern model architectures and attention mechanisms
- Experience with batching, scheduling, and concurrency control in inference systems
- Familiarity with KV cache management and memory placement strategies
- Experience profiling and tuning latency- and throughput-critical systems
- Software development experience in Python and C++
