Meet Dev Shah

Summary of Dev Shah’s portfolio website

Dev Shah — also known as 0xdevshah — is a 24-year-old technologist from Mumbai, India, currently based in San Francisco. He describes himself as a builder and writer with expertise in statistical modelling, agentic systems, complex UX problems, and recreational mathematics.

Dev currently works at Resemble AI, where the team builds voice models and deepfake detection systems. His site says Resemble has raised a $13M Series A and frames the work culture as fast-moving: “we break things, we move fast and, we build faster.”

His portfolio site, omnificence.xyz, is primarily a personal homepage plus an essay/blog archive. He notes that the site is still a work in progress and that although he used to write essays more regularly, most of his time is now spent working on Resemble. The site also includes an Engineering section focused on applied agentic engineering.

What he works on

Dev’s technical interests are centered around AI systems, especially practical engineering for agents. His engineering essay, “The Glass-Box Framework: Reasoning Observability for AI Agents”, argues that production AI agents fail because teams cannot see inside the reasoning loop. He proposes a set of practices for making agents more transparent and debuggable, including structured intent deconstruction, visible reasoning chains, evidentiary citations, epistemic boundary reports, and “axiomatic lens rotation.” The post reflects hands-on experience building deep research agents and a strong interest in moving agent work from demos into reliable production systems.

His broader essays show recurring themes:

• AI-native research and discovery engines — In “How to Invent Everything at Once”, he argues that science is shifting from problem-driven research to method-driven research: general-purpose “discovery engines” for fields like biology, therapeutics, materials science, and semiconductor design.
• Writing and UX as cognitive engineering — In “Narrative as User Experience Design”, he treats writing as a systems-design problem: good prose manages the reader’s cognitive load like a cache.
• Builder taste and specificity — In “The Specificity Principle”, he argues that hyper-specific, hard-won details are more persuasive than abstract claims because they prove real contact with the problem.
• Technology, philosophy, and spirituality — In “Some Things I Believe”, he writes about biocomputing, agentic-predictive interfaces, BCI, VR, determinism, meaning-making, techno-spirituality, Jain influence, craft, learning, and fitness.

Personal interests and worldview

Dev presents himself as both intensely technical and philosophically inclined. He says he is influenced by Jainism and the teachings of Lord Mahavira, and he writes about determinism, “optimal will,” meaning-making, humility, and the importance of changing one’s mind with new evidence.

Outside of work, he mentions lifting weights, boxing, sketching, cooking, and writing. He rejects a strict work-life balance framing in favor of integration and “harmony.”

Social profiles

The site links to Dev’s profiles on X, LinkedIn, and GitHub.