Short bio: I am an AI Engineer at Mantic, where I work on building AI systems for forecasting future events. Previously, I was an AI Engineer at FlowX.AI, where I used LLMs to automate banking app development, and I conducted research with the Visual Geometry Group at the University of Oxford on text-video retrieval.

Research Highlights

AI Forecasting

• At Mantic, I work on automating judgmental forecasting across geopolitics, business, policy, technology, and culture.
• Our systems are inspired by human superforecasters, but aim to deliver predictions at unprecedented accuracy and scale.
• See the launch post for more on how we’re pushing the frontier of AI forecasting.

AI Content Fact-Checking

• At Filtir, focused on fact-checking AI-generated content.
• Built algorithms for cross-referencing and validating data against trusted sources to combat misinformation.

Text-Video Retrieval

• Researched the limitations of text-video retrieval methods.
• Introduced hubness mitigation strategies and developed systems achieving leading results on benchmarks.
• Collaborated with the Visual Geometry Group at the University of Oxford.

Object Segmentation and Detection

• Worked on unsupervised learning, specifically transferring knowledge from video to images.

Engineering Highlights

FlowX.AI

• Worked as an AI Engineer using LLMs to automate the creation of complex banking applications.

MariaDB

• Led the creation of MariaDB’s automated testing framework.
• Ensures compatibility and testing across multiple platforms and operating systems for the MariaDB Server.