Amit Prusty

I build production-grade agentic AI systems that are evidence-backed, measurable, and trusted in enterprise workflows.

Featured Build: Agent Bridge

Local-first CLI for evidence-backed cross-agent context handoffs.

• Repo: https://github.com/cote-star/agent-bridge
• Problem: multi-agent workflows break when handoffs are memory-based and unverifiable.
• Approach: one agent reads another agent's work with citations and structured evidence.

What I Focus On

• Production multi-agent systems and orchestration reliability.
• LLM and VLM engineering with research-to-production translation.
• Full-stack AI engineering (model, retrieval, infra, observability).
• LLMOps, evaluation, and governance for enterprise deployment.
• Strategic AI implementation tied to measurable business outcomes.

Writing

I publish practical field notes on enterprise AI reliability and agent systems:

• Substack: https://thoughtfulengineerconfesssions.substack.com/
• The Silo Tax: https://thoughtfulengineerconfesssions.substack.com/p/the-silo-tax
• Why Agents Don't Fail Fast: https://thoughtfulengineerconfesssions.substack.com/p/why-agents-dont-fail-fastthey-fail

Engineering Principles

• Evidence over assumptions.
• Reliability over demo polish.
• Measurable outcomes over vague AI claims.
• Simplicity first; orchestration only when needed.

Connect

• LinkedIn: https://www.linkedin.com/in/amitprusty/
• GitHub: https://github.com/cote-star
• Substack: https://thoughtfulengineerconfesssions.substack.com/