How this works

AI assistant for Dhavan’s published portfolio material. Answers can be incomplete or wrong, so do not submit confidential code or data.

Your questions go through Hugging Face, and a model run by DeepInfra writes the answers. Chats may be retained for about 30 days to operate, debug and improve the assistant; DynamoDB TTL deletion is approximate. Cal.com handles call bookings. Details are in the privacy policy.

Use the suggested questions or ask about published projects, credentials, and engagements.

Projects / Agent Builder

A coding agent that asks before it acts.

Agent Builder is a plan to code workflow with approval gates, spend logs and a VS Code extension, so a developer reviews the risky steps before they land.

Dhavan Shah You work with the person who built it. The first 30 minute call is free, and if I can’t help, I say so before you pay anything.

The gatesBy design
Plan
You approve
Proposed changes
You approve
Spend
Logged
Where you work
VS Code

AvailabilityInternal tooling

Assisted coding can spend and change code before anyone has approved it.

Plan checkpoints, approval gates and spend logs slow only the steps that need a person’s authority.

  • “Can it change code before I’ve looked?”

    Approval checkpoints

    A developer reviews plans and proposed changes before accepting them.

  • “What is it doing right now?”

    Live run visibility

    Progress, status and cost stay readable during the run.

  • “Do I have to leave my editor?”

    Inside your IDE

    Prompts, progress and change review stay close to daily engineering work.

  • “What happens when my guidance changes?”

    Feedback loops

    The system can revise its approach instead of failing silently.

Slow only the steps that need authority.

How it works: plan, approve, review.

A workflow from task to reviewable output, instead of one shot generation.

1 · Plan

A plan you can read

The task becomes a plan a developer can check before any work starts.

2 · Approve

A person holds the gate

Risky implementation and cost moments wait for human review.

3 · Review

Accept or send it back

Inspect proposed work before it’s final, or give guidance and let it revise.

What it proves.

A portfolio case study of a governed coding workflow.

  • Developer agents can stay useful while requiring human review at high risk implementation and cost moments.
  • Coding agents become more practical when orchestration, approval, visibility and developer experience are designed together.

Where it’s useful

When teams want agent assisted delivery without giving the agent silent permission to execute everything.

Availability

Internal tooling I adapt in engagements, not a product you buy. Use it as a pattern to evaluate human gates in your own agent workflow.

Tell me where your agents shouldn’t act without a person checking first.

The first 30 minute call is free, and I reply within one business day.

Book a free call

How engagements work · Back to Projects