Tech Career Compass
LiveAI-powered career coach that helps digital professionals find suitable career paths and build a concrete study plan
Starting Point
Digital professionals live in the middle of constant change. AI is reshaping job descriptions, new technologies appear monthly, and role boundaries are blurring. In an especially challenging position are people with skills across multiple areas — some analytics, some marketing, some technical understanding — but who aren't top experts in any single field.
For these people, traditional career services work poorly. LinkedIn suggests the same roles as their current job. Generic career guides advise "specializing," even though versatility is these people's strength. And the real state of the job market — what roles are emerging, what skills they require — is scattered across dozens of different sources.
Why This Was a Problem
Without a clear picture of their options and requirements, people easily get stuck in their current role or apply randomly. They might invest months studying the wrong certification when the relevant skill was elsewhere. Or they don't see that their existing skill combination — for example, data analytics + business understanding + basic coding — is exactly what certain growing roles require.
The problem isn't lack of information. The problem is that no one connects an individual's situation with market realities at a personal level.
How I Approached the Problem
I started from my own experience. I'm exactly that "multi-skilled person without one clear specialization" — background in e-commerce, analytics, programming, and digital marketing. I knew what it feels like when you don't know how to position your skill set.
From my previous Shopify app project (Portfolytics), I had learned that you shouldn't build a product to completion before validating demand. So this time I started faster: first a small prototype, then real users testing as early as possible.
I also researched existing solutions. Career coaching apps exist, but they typically offer either generic personality tests or static content libraries. None combined real-time market data with an individual's skill profile to produce a tailored action plan.
What Was Built
The app's core idea is that a user gets a personalized picture in a few minutes of what roles they should consider and what those roles require.
The user journey starts with two simple questions: what's your current situation and what motivates you. No registration, no CV upload, no lengthy survey. After that, the app uses AI and real-time market data to produce a tailored view of possible roles — not a static list, but analysis-based recommendations.
When a user selects an interesting role, the app does a skill assessment and identifies missing skills. On top of this, it builds a study plan: concrete steps, daily tasks, and progress tracking.
A key design decision was that value is delivered immediately before asking the user for anything extra. Earlier versions started with extensive data collection — skill assessments, background information, goal setting. Testing showed this killed initial enthusiasm. The current version gives the first "aha moment" in seconds and deepens data collection only after the user has seen what the app can offer.
Technical Implementation
The app is built with web technologies on the Lovable platform and currently works as a PWA (Progressive Web App). Content isn't hardcoded — the app uses LLM APIs and web search to produce real-time, personalized content. This means recommendations always reflect the current state of the job market and don't go stale.
Result
The product is live and free to use. There's potential for monetization if I could get help with marketing. If you're a strong marketer and want to join the project or collaborate, reach out!
What I Learned
First impressions are everything on mobile
The original user journey asked too much too early. Testing clearly showed that a user needs a reason to continue first — not a form to fill out. When I flipped the order so that the personalized view comes before deeper data collection, the user experience improved significantly.
This is the same lesson as in the Portfolytics project: talk about the benefit, not the features. "AI-powered skill assessment with real-time market data" is a technical description. "See in two minutes which roles your skills qualify you for and what you should study next" is a promise that makes a user continue.
Hardcoded content goes stale, API-powered doesn't
An early version contained static role descriptions and skill lists. The problem is that digital roles change constantly — a "relevant skills" list from six months ago can already be outdated. Using LLM and search APIs solves this but brings its own challenges: loading times, uncertainty in responses, and the increased importance of UX design during waiting states.
Learning from previous projects is the most valuable thing
Portfolytics taught me not to build a product to completion before validation. In this project, I applied that lesson — the first test users came on board much earlier. Every project builds on the previous ones, and that's perhaps the most important reason to do multiple projects: not just for the end results, but for the cumulative learning.