SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Fireworks is hiring a Member of Technical Staff to own the evaluation and post-training product infrastructure. The role sits at the intersection of scalable system design, data science, and model quality—building the infrastructure and workflows that connect evaluation and post-training for developers improving models on the Fireworks platform.
Key responsibilities include scaling evaluation infrastructure by designing systems to eliminate coordination bottlenecks and resolve log-syncing latency; acting as a "metrics obsessive" to track state-of-the-art benchmarks, read research papers, and rigorously reproduce published results; designing and building new benchmarks for complex, emerging, or domain-specific use cases; owning user-facing workflows for post-training including fine-tuning experiences across SFT, RFT, and related capabilities; and partnering with customers and internal stakeholders to understand needs and convert bespoke workflows into productized solutions.
The ideal candidate has 1–7 years of software engineering or data science experience (the role is open at multiple levels), strong system design skills for architecting scalable, programmatic systems, hands-on experience with sandbox infrastructure for secure code execution, deep understanding of LLM evaluations and how to design them, and understanding of the end-to-end GenAI lifecycle from prompting to fine-tuning to productionization.
Preferred qualifications include 3+ years of software engineering or applied data science experience, familiarity with container registry and orchestration tools, interest in public technical writing and analysis, knowledge of GPU constraints and inference optimization, and startup experience with end-to-end feature ownership.
Fireworks is a Series D company valued at $17.5 billion, founded by the team behind PyTorch and backed by AMD, Atreides, Benchmark Capital, Index Ventures, Lightspeed, NVIDIA, Sequoia Capital, and TCV. The company powers production AI with hundreds of state-of-the-art open models across text, image, embedding, audio, and multimodal workloads.
