Emilio Andere

Founder & systems‑level ML engineer focused on inference performance. Emilio co‑founded Herdora (YC S25) to tackle production ML inference — building tooling that traces from PyTorch down to the GPU, generates/optimizes kernels, and offers routing/autoscaling and observability. The Herdora repo (kandc / Keys & Caches) and YC/company listings show this emphasis on kernel- and stack-level performance. ycombinator+2

Research background in ML security and adversarial/backdoor methods. As an undergraduate researcher in UChicago's SAND Lab, Emilio co-authored the NeurIPS/arXiv work 'Finding Naturally-Occurring Physical Backdoors in Image Datasets' and contributed to the lab's codebase for generating those datasets — demonstrating strengths in experimental ML, dataset construction, and adversarial-robustness research. arxiv+2

Bridges research and product engineering. Public materials and profiles describe a trajectory from hands-on academic research (UChicago / SAND Lab) into short engineering roles at multiple startups and then founding a YC-backed company; this pattern indicates an ability to move from research prototypes into production software and productize systems-level ML research. uchicago+2