INSIGHTS

Obin Blog

Obin Blog

Research, product notes, and perspectives on building AI agents for regulated finance.

Engineering

Jul 16, 2026

Plausible vs. Optimal: The Danger in using LLMs to write Investment Memos

Engineering

Jul 1, 2026

How Regulatory Scrutiny of AI Vendors Is Changing Procurement

Financial regulators are reshaping enterprise AI purchasing decisions – here's why and what organizations must do to satisfy evolving governance expectations.

Design

Jun 4, 2026

The UX of AI Applications depends on the Harness, not the Model

Most AI products succeed or fail based on orchestration, controls, and trust-building mechanisms that shape behavior, reliability, transparency, and user confidence.

Engineering

May 27, 2026

You're About to Become a Manager of Agents

This is part two of a two-part blog adapted from my conversation with Marc Andrew on Modern Capital: The Private Markets Podcast.

Engineering

May 21, 2026

What Leading AI at Google, JPMorgan Chase and Silver Lake Taught Me About AI Transformation

This is part one of a two-part blog adapted from my conversation with Marc Andrew on Modern Capital: The Private Markets Podcast.

News

May 8, 2026

The AI Last Mile: Anthropic Just Made the Case for Embedding Forward-Deployed Engineers in Financial Services

A look at why embedded AI consulting models may recreate the same scalability and lock-in problems financial institutions have faced with legacy enterprise software and system integrators.

Engineering

Apr 29, 2026

Moving Document AI from Experimentation to Engineering

Obin AI’s Context Studio transforms document AI from fragile experimentation into a structured, version-controlled engineering workflow for financial institutions.

Engineering

Apr 15, 2026

Future Bankers Build AI Workers at Cornell Workshop

A Cornell workshop shows how AI is rapidly reshaping finance careers, forcing the next generation to shift from doing analyst work to building and overseeing intelligent systems.

News

Apr 15, 2026

Obin AI Announces New Collaboration with Google Cloud

Obin AI’s platform enables financial institutions to deploy a production-ready agentic workforce on Google Cloud

Engineering

Mar 31, 2026

Addressing LLM Availability Constraints with Proactive Congestion Control

Obin improved on a reactive AIMD rate limiter with a hybrid latency-adaptive strategy. When making LLM calls at scale, proactively adjusting concurrency based on API response times achieved 7x throughput and 98% fewer errors in our distributed document-processing pipelines.

News

Mar 18, 2026

Obin AI Emerges from Stealth to Build the AI Workforce Financial Institutions Can Actually Trust

Backed by Motive Partners and AI pioneers Dr. Fei-Fei Li and Lukasz Kaiser, Obin AI is closing the gap between AI potential and regulated deployment — already trusted by firms managing over $1 trillion in assets under management.

Engineering

Jan 11, 2026

The platform and the people you need to build Vertical AI

Why do most AI pilots in finance fail while a few actually scale? This post uncovers the hidden infrastructure and domain expertise needed to move from impressive demos to reliable, production-ready AI, and explains why finance demands far more than “mostly correct” models or better prompts.

News

Jan 1, 2026

Why did we start Obin AI?

The founding story of Obin AI

News

Hard-Won Lessons from Deploying AI Agents in Regulated Environments

Top takeaways from an exclusive NY Tech Week panel with operators from JPMorgan Chase, Apollo, and Google Cloud on building enterprise AI that ships, scales, and survives a regulator's review.

Obin AI

Obin builds specialized agentic workforces for regulated financial institutions – rooted in your institutional knowledge and designed for the edge cases where most AI stalls

Obin AI

Obin builds specialized agentic workforces for regulated financial institutions – rooted in your institutional knowledge and designed for the edge cases where most AI stalls

Obin AI

Obin builds specialized agentic workforces for regulated financial institutions – rooted in your institutional knowledge and designed for the edge cases where most AI stalls