Skill Tree

Simulate your career before you commit — model skill growth, salary projection, and market risk to choose the smartest path forward.

Comment

Inspiration

Most people still make career decisions based on trends, opinions, or social proof, not structured analysis.

We wanted a tool that lets someone test a career move before spending months (or years) learning the wrong things.

SkillTree was built to push career planning from guesswork toward data-informed strategy.

What it does

SkillTree is an AI-assisted career path simulation dashboard.

Users enter:

  • current role
  • years of experience
  • current skills
  • desired role
  • location
  • risk tolerance
  • time horizon (1-5 years)
  • weekly learning hours

SkillTree then generates and visualizes:

  • A 6-month learning/action roadmap (with resources and certifications)
  • A multi-year salary projection (1-5 years)
  • Transition probability over time
  • Market metrics (demand, competition density, volatility, saturation, automation risk)
  • Risk breakdown (obsolescence, automation, volatility, saturation trend)
  • Skill investment ROI estimates (salary impact / demand increase / risk delta / effort hours)

Users can also:

  • Compare saved simulations side-by-side
  • Export a PDF report
  • Load role news/trend summaries with sources
  • Edit the underlying skill graph in a built-in admin UI

Instead of only asking "What should I learn?", SkillTree helps answer:

"If I learn this, what could happen to my career in the next few years?"

How we built it

  • Frontend: React + Vite + Recharts + Lucide
  • Backend: Node + Express API
  • AI layer: Gemini (@google/genai) for roadmap/simulation/news/action suggestions
  • Skill graph: SQLite (better-sqlite3) with CRUD/query endpoints
  • In-app Graph Admin: create/edit/delete/import/export JSON/CSV for skill graph edges
  • Skill optimizer: backend endpoint using curated skill ROI signals + role-aware graph adjustments

The current system blends:

  • AI-generated planning/simulation outputs
  • a local persisted skill graph (shared by simulation + optimizer)
  • curated skill signal data for some ROI estimates
  • heuristic fallback scoring where full datasets are not yet integrated

Challenges we ran into

  • Moving from a frontend-only prototype to a backend API architecture (to avoid exposing API keys)
  • Defining simulation outputs that feel useful while staying honest about uncertainty
  • Designing a reusable skill graph that can influence both simulation and ROI optimizer logic
  • Building a graph admin workflow that is fast to iterate on during a hackathon

We addressed this by:

  • introducing a backend proxy/API layer
  • surfacing confidence intervals and metric breakdowns in the UI
  • keeping a clear separation between curated data, graph signals, and heuristic fallbacks

Accomplishments that we're proud of

  • Built a working multi-scenario career simulator
  • Added multi-year projections plus risk/market breakdown visualizations
  • Built a SQLite-backed skill graph with live in-app admin tools
  • Implemented a skill investment optimizer with curated + graph-adjusted scoring
  • Shipped an end-to-end backend API + frontend dashboard flow

What we learned

Career growth is multi-dimensional.

Salary alone is not enough — risk, demand, competition, and skill longevity all matter.

We also learned that predictive UX needs transparency:

  • users trust the output more when they can see assumptions, breakdowns, and confidence bands
  • an editable graph/admin UI is valuable even in early demos because it makes the system feel inspectable

What's next for SkillTree

  • Real job market / salary datasets (reduce heuristic placeholders)
  • Confidence intervals backed by real compensation distributions
  • Skill graph persistence + management improvements (auth, versioning, audit trail)
  • LinkedIn / GitHub profile import
  • Personalized AI career mentor
  • Team talent planning / workforce intelligence dashboard

SkillTree can evolve from a personal career simulator into a broader workforce intelligence platform.

Built With

Share this project:

Updates