March 13, 2019

[Feedback] AI Career Advisor

Hi all! I've been working on PathBase, an AI career assistant that helps young professionals build dream careers by spotting their biggest skill gaps, and bringing top resources (on the web & at your company) to develop those skills.

This project started from my own frustrations of being in a job unaligned with my passions, with limited time/energy to grow on the side, and a super-obscure career development process.

See it here: https://pathbase.io (beta code = 1nd13h4x)

Would you use this? Is it missing anything? I could really use some feedback!

#product-feedback-request #idea-validation

  1. 2

    Isn't career builder similar? But yeah, there is definitely demand for this.

    1. 1

      Thanks @genemachine. 🙌 Perhaps we're talking different "career builder"s, but I'd the say the biggest difference is that while Career Builder is more of a job board that connects you with jobs, PathBase prepares you to be a strong candidate for those job opportunities.

      A good analogy would be like a Personal Finance app that helps you with your financial goals (or in this case, Career goals). 🎯

      Would you mind expanding a bit on which aspects you think is demanded? 🙇‍♂️

  2. 1

    Says, beta closed? Can I try this out? Will appreciate invite!

    1. 1

      Absolutely! I actually just extended the closed Beta date a few days (open Beta coming next week), so would love to have you try it!

      Use this link to sign up: https://pathbase.io/signup/1

      Beta Key: 1nd13h4x

      1. 2

        Thanks! Will try it out. Btw, where do you see this going? Can you elaborate more on your intents, goals and what will be milestones ?

        Business model?

        1. 1

          Hey @madebymaya, thanks for asking - happy to share!

          So far, the focus was on getting the individual's experience right. This means that the app is able to quickly analyze a user's resume/profile, derive where they are with each skill in their skill set, and make recommendations on which skills to build and how.

          I believe the project is very close to the above (in fact, public Beta is launching in a week!). The next step is to enable collaborative learning by connecting users to each other (via teams and/or public communities) - features that will enable the business model (below).

          The final goal, which ties to the business model, is enabling companies to optimize employees' lifelong learning so the skills they build are aligned both with their own career goals, as well as the companies' skill gaps.

          Most companies struggle with retention & unfilled talent gaps, and most people don't use their companies' existing Learning Platforms (too generic in recommendations). By aligning projects & microlearning to employees' career goals, companies can become "self-healing" in addressing their talent needs.

          You can read more about the roadmap (here)[https://medium.com/pathbase/pathbase-master-plan-optimize-career-building-pathbase-blog-963bdbf5fd79?source=your_stories_page---------------------------]. Let me know if you have any more questions and/or comments!

          1. 1

            Thanks. Have signed up for beta and checking out the platform. The MasterPlan is good!

            What's the underlying stack? And the AI analyser / advisor itself is based on ? Is there enough dataset available to derive such analysis (lots of AI powered resume analysis out there already but am kinda wary about them).

            1. 1

              Hey @madebymaya, thanks a ton for checking it out! Really means a lot for me & the project. Btw, we just released Open Beta (https://www.producthunt.com/posts/pathbase), so you may want to check it out now. We're also wrapping up a few major updates on the accuracy & breadth of our resource recommendations, so stay tuned!

              As for your questions:

              1. The underlying stack is ReactJS + Firebase (Node.js) on the back-end.

              2. AI Advisor is based on job posting data, where we associate weights between every career path & skills (our highly standardized & curated skills database is at 700 and growing).

              This is what drives the recommendations for everything in the app, from recommending resources (e.g. articles, courses), to helping users figure out what their biggest skill gaps are.

              Obviously far from perfect yet, but definitely yielding very strong recommendations for certain users (we've focused on nailing the most popular Paths first).

              Lmk if more questions! Have you tried the app? Dying to hear your feedback =)

              1. 1

                Thanks a lot for your detailed response. Did check out the PH launch page. The Path recommender is working well for the popular paths (Data Science, Python etc) & has plenty of relevant articles, courses etc. Congrats on getting to these nice levels.

                Surely it will improve more as time goes and with more users / predictions. As with any deep paths (like engineering etc) there is not much to offer at this point but surely bound to improve. Is there any plan on tracking how the user learning progresses owing to all these recommendations?

                Also, on the business model itself, as it's free for users & only paid for businesses, do you have any particular number of different client acquisitions in mind (users / businesses)?

                1. 1

                  Hey @madebymaya, sorry about the tardy response. Got overwhelmed with the Open Beta, integrating user feedback and new features, leading me to lose track of things... (I'm also doing this while in a full-time job to maintain VISA, so hope you understand...!)

                  I'm really happy to hear you found the Path recommendation engine pleasing! On that point, yes absolutely, it will improve in the near future with deep data dives in the queue, and eventually users & predictions (as we integrate ML). The reason you don't see robustness in some of the "deep paths" (e.g. engineering) is due to my hypothesis on the early adopter segment - "hybrid" paths (e.g. product management, front-end), which have less "deep" educational requirements. But I'm quickly finding that folks in engineering also face a similar challenge, especially as they approach their 2nd-3rd years in their careers.

                  On your point on "tracking user learning", that is, in fact, one of the key UX overhauls I'm doing atm, since we want to provide users an accurate sense of progress as they consume content, complete courses and projects at their job. The ultimate goal is to make it feel like an MMORPG, where people would gain experience points for the above activities (just like you get points for slaying monsters & completing quests). It's a critical feature that'll (a) boost user engagement via gamification, and (b) enable powerful organizational learning & skill analyses.

                  Now that it's in Open Beta, I wanted to validate the organizational value propositions by piloting the upcoming features (learning playlists, team spaces, organizational learning analytics - most of which are in the final QC stages) this with around 5-10 small-medium-sized companies The hope here is that this not only validates the customer need but also boosts user retention via collaborative learning.

                  From what you've seen so far of the app, what do you think is missing (if any) to get you to use PathBase on a regular basis? I would love to integrate your feedback before the next big update in around a week!

              2. 1

                Thanks a lot for your detailed response. Did check out the PH launch page. The Path recommender is working well for the popular paths (Data Science, Python etc) & has plenty of relevant articles, courses etc. Congrats on getting to these nice levels.

                Surely it will improve more as time goes and with more users / predictions. As with any deep paths (like engineering etc) there is not much to offer at this point but surely bound to improve. Is there any plan on tracking how the user learning progresses owing to all these recommendations?

                Also, on the business model itself, as it's free for users & only paid for businesses, do you have any particular number of different client acquisitions in mind (users / businesses)?

  3. 1

    Personally my brain tunes out whenever I see "crypto" or "AI". IBM still couldn't quite figure out their Watson stuff. How do I trust this service? I would be more inclined to buy in if it's worded with deep learning/machine learning. Or perhaps you could somehow incoprorate personality traits (ex Briggs Myers) to the mix? Hope my feedback doesn't come off harsh. Cheers!

    1. 2

      I agree here. AI is too lightly thrown around. But the core idea is still great IMO. What you'd need is to be transparent about how it works and most importantly the benefits it brings to the user s (companies, and applicants).

      1. 1

        Thanks again @genemachine! I agree on the transparency bit, and am actually working on incorporating it into the recommendations (i.e. "We're recommending you this because of X, Y, Z").

        As the goal with this is to expand into companies as a Corporate Learning Experience platform (see here) that enables employees to constantly learn collaboratively. 🎓

        I'm actually refining the landing page for businesses to test with teams, so it would be incredible if I could get your thoughts on it! 👂

    2. 1

      Hey @indiepanda, I totally agree with the whole buzzword thing, which (as @genemachine says) is thrown too lightly. This actually is a recent change in messaging after trying different ways to better express the idea concisely.

      Yes, the app is trending towards full-on AI capabilities (as things scale), but the bigger reason I used "AI" is to convey that this app is like an entity dedicated to your career and will really dig in to where you're at with your skill set to make very granular recommendations (which skills you should be focusing on, and content & resources to learn those specific skills). 🧐

      Would you have a better way of expressing the value prop without leaning on the "AI" word? 🤔

      As for your latter point, incorporating personality trait standards like Myers Briggs is definitely in the pipeline, and agree it will help pinpoint the type of career the user wants to build. Probably will start on this once we lock down the accuracy of the recommendations (e.g. articles, videos, podcasts, courses) down to a decent point. Please lmk if you have more ideas on what "magical feature" this may be missing! 🙏