Skip to content

Career Fit Navigator

Imagine us
working together.

Bring a role. Explore the work behind it. Leave with better questions.

What are you building?

Give Jev a role to interpret. See which parts connect to my work.

184 / 8,000 characters

Use public, non-confidential text. The application does not save your description. Live interpretation sends your text to Jev through Vercel AI Gateway; the provider’s data policies apply. If free access is unavailable, the result is labeled as a keyword preview.

Your conversation starter.

Illustrative example · no AI call

This preview uses keyword matching. It can miss negation and unstated requirements. No live Jev interpretation was performed.

AI agentsRetrieval & RAGBackend & APIsIntegrationsProduct delivery

01 / Loan Labs

From agent APIs to a product interface

Built agent-facing Rails APIs, borrower email intake, and an in-product conversational interface so users could handle loan documents and initiate agent actions in LoanOS.

Related to Product delivery · Backend & APIs · Integrations

Internal and pilot workflows; no claim of broad customer rollout.

Read the public work

02 / Loan Labs

Lois: agents for mortgage workflows

Re-architected Lois from one-off Ruby LLM calls into a LangGraph agentic system on Amazon Bedrock AgentCore for mortgage-document classification, lender-specific renaming, and policy validation.

Related to AI agents · Backend & APIs

Internal and pilot workflows; no claim of broad customer rollout.

Read the public work

03 / Loan Labs

Permission-aware agent actions

Designed fail-closed authorization for Composio integrations: tenant/owner scoping, separate write/send/merge/archive permissions, reviewed tool allowlists, and execution-time checks that invalidate revoked access.

Related to AI agents · Integrations

Internal and pilot workflows; no claim of broad customer rollout.

Read the public work

Ask me about

  • How would you decide which actions an agent can take, and which need review?
  • What would you ship first, and what evidence would change that decision?
  • How would you evaluate retrieval quality before adding more model complexity?
Start the conversation