Summary
Christian Cosgrove is a Senior Software Engineer in the San Francisco Bay Area with 12 years of experience building production ML systems spanning computer vision, adversarial robustness, and large-scale generative language models. He has driven safety and evaluation work at Microsoft for Bing Chat/Copilot, led research and product engineering at Solver to top performance on SWE-Bench, and now contributes at NVIDIA. His background blends academic research—novel adversarial attacks/defenses and contrastive learning—with hands-on infra for training, sharded eval pipelines, and product integrations like Slack and voice interfaces. He has applied ML in safety-critical domains at Tesla (pedestrian detection, AEB) and repeatedly shipped features that moved metrics and deployments forward. Equally comfortable in low-level scientific computing, native apps, and web stacks, he pairs rigorous research instincts with a founder’s appetite for shipping. A subtle throughline: he often turns evaluation and tooling challenges into competitive product advantages, e.g., sharding thousands of benchmarks to accelerate iteration.
12 years of coding experience
6 years of employment as a software developer
New Jersey Governor's School in the Sciences
Johns Hopkins University
Valedictorian, Valedictorian at Ramapo High School