Nemora

Nemora is an AI copilot that turns a PM’s idea into an execution-ready roadmap in seconds.Powered by NVIDIA Nemotron, it creates phases, tasks, risks, metrics, and assignees.

Comment

Nemora – Product Lifecycle Copilot

Turn a product idea into an execution-ready lifecycle in one click.

Our Inspiration Behind Nemora

We built Nemora directly from the PNC challenge statement.

Their prompt around helping Product Managers generate structured artifacts (stories, plans, etc.) clicked with our own experience: PMs spend a lot of time turning complex ideas into documents, tickets, and timelines. Most of this is mechanical and repetitive.

At the same time, NVIDIA Nemotron was sitting there as a powerful LLM made for exactly this style of structured generation. It felt like a natural pairing:

  • PNC → real PM workflow & structure
  • NVIDIA → model that can generate high-quality, controllable plans

Nemora is the intersection of those two: a focused place where PMs can quickly go from “we should build X” to a concrete lifecycle they can use with their teams.

What Nemora does

1. Sign in & set context

PMs log into their account so their work lives in one place (no anonymous one-off pages).

2. Describe your product

One short form:

  • Product name
  • Short description
  • Target users
  • Timeline

No giant questionnaire. Just enough info for a smart default plan.

3. Generate lifecycle plan (Nemotron-powered)

On Generate Lifecycle Plan, our backend calls Nemotron with a structured prompt.

Nemora returns a lifecycle split into phases like:

  • Discovery
  • Design
  • Build
  • Test
  • Launch
  • Post-launch

4. Execution view for PMs

From the lifecycle, Nemora also generates:

  • Tasks grouped by phase
  • Priority levels (P0/P1/P2)
  • Status fields PMs can update

The idea is: PMs can use this as a starting point for their real execution backlog.

5. Risks & metrics

Nemora suggests:

  • Potential risks (technical, regulatory, operational)
  • KPIs that match the product and timeline

This is especially aligned with the PNC/finance use case where risk and measurement are critical.

How we built Nemora

Frontend

  • React + TypeScript (.tsx)
  • Tailwind-based, dark theme inspired by tools like Linear

  • Sections:

    • Auth / login
    • Product input form
    • Lifecycle overview
    • Execution plan
    • Risks & metrics

Backend

  • JavaScript server / API routes

  • Handles:

    • Auth session
    • Validating product input
    • Calling Nemotron safely using server-side API keys
    • Normalizing responses into a consistent JSON structure consumed by the frontend

Why this fits sponsors

  • PNC: Focused on real PM workflows, structured outputs, and lifecycle planning.
  • NVIDIA: Uses Nemotron as the engine; not as a chatbox, but as part of a multi-step planning workflow.

What we learned

  • Tightening the schema (phases, tasks, risks, KPIs) massively improves AI usefulness.

  • LLMs work much better when we:

    • clearly define fields,
    • give examples,
    • and always expect structured output.

  • It’s easy to overbuild:

    • We initially thought about full project management,
    • We cut it down to one strong flow: log in → describe → generate → refine.

  • Working with Nemotron pushed us to think about prompt design like API design, not just “ask for text.”

Challenges we faced

  1. Keeping it structured, not chatty

Our first attempts produced good content but unstructured blobs. We had to iterate prompts and response handling until Nemotron reliably returned data we could map into UI components.

  1. Balancing scope with polish

We wanted:

  • auth,
  • clean UI,
  • meaningful AI output.

In hackathon time that meant being ruthless:

  • no heavy integrations,
  • no complex multi-tenant dashboards,
  • just the core PM journey done well.
  1. Making the UI feel “real”

Getting the layout, spacing, and dark theme to feel like a modern PM tool (instead of a quick demo) took more iteration than expected, but it sells the idea.

  1. Safe & correct model usage

Making sure:

  • Nemotron keys live only on the backend,
  • calls are efficient,
  • and timeouts/failures degrade gracefully.

What’s next

We want to keep working on Nemora, and integrate the following:

  • Persist multiple product lifecycles per user.
  • Export straight into Jira/Linear or CSV.
  • Add domain templates (e.g. payments, lending, AI infra).
  • Collaborate in real time across PMs and teams.

For this hackathon, Nemora is our focused proof-of-concept: a simple login, a short description, and a serious AI-generated lifecycle that PMs can actually use.

Share this project:

Updates