Seenia Hong

Seenia Hong works on AI deployment, with experience in technical strategy, partnerships and data-product management. She built partnerships and the technical ecosystem at Edison Scientific, and her personal website describes her current work as special projects with early-stage technical founders in San Francisco. Her stated interest is what bottlenecks frontier AI deployment: data and compute, economics, commercial structures and the incentives that determine what ships.[2]

In a March 19, 2026 announcement, Hong said she had joined Edison Scientific to lead Technical Strategy and Partnerships for its intelligence stack. She framed its purpose directly: “The AI and science frontiers don't intersect enough, and we're here to change that.”[4]

Hong coauthored “Accelerating Science at Scale”, published March 16, 2026, with Jon Laurent, Siddharth Narayanan and Richard Magness. They described Edison's collaboration with NVIDIA on training infrastructure and evaluation for scientific agents, using NeMo and Nemotron. One concrete deployment problem was preserving scientific evidence during document retrieval: figures, tables, formulas and captions can carry findings that an abstract omits. The article explained how Nemotron Parse processes pages into classified bounding boxes and parsed text to preserve that structure.[3]

The same article distinguished correct analysis under explicit instructions from analytical judgment under ambiguity. Its BixBench-Hypothesis benchmark pairs data, a hypothesis and an analysis protocol with a rubric that scores intermediate outputs and whether the agent correctly supports or rejects the hypothesis. This was a collective account of Edison's work; Hong's identified responsibilities were strategy and partnerships.[3][4]

Hong, Abhay Venkatesh and Michael Tan co-hosted the AI x Healthcare industry dinner group. Their event description argued, “We believe the more immediate gain is to focus on distributing AI.” The proposed format brought industry insiders together with technology and capital providers to discuss specific needs and deployment bottlenecks, in groups of eight to twelve people.[6]

Hong also co-hosted Latent Space LIVE! at NeurIPS 2024, which she describes as a one-day paper survey conference. Latent.Space's 2025 AI Engineer World's Fair recap credits her as the AI PM track host.[2][1]

Before Edison, Hong's work included commercial responsibilities for data products and AI research. She describes being the business lead on an Apple AI/ML R&D team and owning profit and loss for a telemetry-inference data product at Mapbox. In 2022, Hong was listed as Mapbox's Product Manager, Telemetry Derivatives, and scheduled for “Data for Good Is Good for Business” at Development Data Partnership Day. She describes the beginning of her career as turning petabytes of data into evidence presented in federal court.[2][5]