Shaped AI has launched a semantic recommendation and search platform for marketplaces and content companies, aiming to simplify the process of building personalized experiences for users.
The big picture: Shaped AI’s platform addresses the growing challenge of content overload by providing tools for businesses to create sophisticated recommendation and search systems without extensive AI expertise.
- The company offers a sandbox environment at play.shaped.ai/dashboard/home for users to explore demo models and evaluate results interactively.
- A demonstration video is available on YouTube, showcasing the platform’s capabilities and user interface.
Key challenges in content discovery: The proliferation of online content, both user-generated and AI-produced, has made it increasingly difficult for users to find relevant information and experiences.
- Platforms like Netflix, TikTok, and Meta have set high standards for personalized content delivery, raising user expectations across the digital landscape.
- Despite advancements in AI and semantic infrastructure, building effective recommendation and search systems remains a complex task for many businesses.
Shaped AI’s approach: The platform aims to simplify the process of creating state-of-the-art recommendation and search systems by addressing key technical challenges.
- The system integrates directly with popular data sources such as Segment, Amplitude, and Rudderstack, streamlining data collection and processing.
- Shaped AI leverages cutting-edge language, image, and tabular encoding models to extract maximum value from structured and unstructured data.
- The platform is designed with real-time optimization at its core, enabling continuous training, feature engineering, and serving of recommendations.
Real-world applications: Several companies are already using Shaped AI to enhance their user experiences and drive business results.
- Outdoorsy, an RV rental marketplace, has implemented Shaped AI to power its recommendation system.
- E-commerce businesses like DribbleUp and startups such as Overlap have reported up to 40% increases in both conversions and engagement after integrating the platform.
Founders’ background: Shaped AI’s founders bring relevant experience from leading tech companies and research institutions.
- Tullie, previously an AI Researcher at FAIR (Facebook AI Research), worked on multimodal ranking at Meta and released PyTorchVideo, a widely-used video understanding library.
- Dan led product research at Afterpay and Uber, focusing on the intersection of behavioral psychology and user experience.
Technical infrastructure: Shaped AI emphasizes the importance of seamless orchestration of various AI components to create effective recommendation systems.
- The platform addresses the challenge of continuously fine-tuning models with fresh data while simultaneously serving real-time personalized recommendations to millions of users.
- By handling complex tasks such as real-time streaming, ETLs, and data quality robustness, Shaped AI allows technical teams to focus on building and refining their recommendation and search systems.
Looking ahead: As Shaped AI launches publicly, the founders are seeking feedback and engagement from the tech community.
- The company is open to technical deep dives, feature requests, and general feedback to further refine and expand its offering.
- This launch represents a significant milestone for Shaped AI, marking the transition from development to public availability and potential widespread adoption.
Launch HN: Shaped (YC W22) – AI-Powered Recommendations and Search