Launch HN: SharpestMinds (YC W18) – Online Community for AI Devs
We're ML developers from non traditional backgrounds. Ed did a PhD in biological physics, and Jeremie studied quantum optics before dropping out of grad school to work on SharpestMinds. We started looking for ML jobs after school, thinking it shouldn't be too hard to get one. We found to our naive surprise that we fell short on a number of skills that are needed to do good work in industry. You just don't learn much devops in grad school.
As a result we decided to build something that would make it easier for ML devs to develop (and discover!) skills they might be missing, and then get their first jobs or internships. From the outset we also wanted to build a community around the process, since looking for your first job is usually a pretty lonely experience. Because we monetize directly through hiring, we can afford to create a space for discussion without ads or algorithmic distractions :)
Our typical users so far have been grad students who know ML material well, but don't yet have much, or any, practical experience. However, you don't need a degree at all (a few of our users are self-taught high school dropouts), and anyone who knows the material is welcome. In fact, that's one of the advantages of our system: we test directly for knowledge, so it doesn't matter how you got that knowledge or how long it took you to get it. One of our goals is that by the time we present you as a candidate, things that would otherwise be holes in your resumé don't matter so much, and we can make that case to companies that are hiring.
To qualify for joining, you do an online deep learning quiz (here: https://www.sharpestminds.com/members/apply), followed by a technical interview. If you pass both, we invite you aboard. It's possible to retake the quiz a month later if you don't pass it, and we'll send you tips on what to study in the meantime.
Once you join you get access to a job board with exclusive (i.e., not scraped) internship and full-time opportunities on it. We've created an application system where your profile gets customized to the job you're applying for, to maximize the odds that you'll get an interview. We also have lists of common interview questions, mentors that you can practice interviewing with, and periodic AMAs with ML hiring managers from companies like Skydio and Airbnb.
The hardest part about building this has been figuring out the best way to present our users to employers. Early on we found that hiring managers were passing on qualified people, because their eyes would glaze over from reading too many CVs. We ended up building application profiles that let our users display their most relevant personal projects prominently in their application. The interview rate has increased significantly as a result.
If our approach works for the ML/AI field, we'd like to build communities like this for other fields too.
We're looking forward to getting feedback and hearing ideas from HN! We know there are lots of ML devs / enthusiasts on here, and we'd also be very interested in hearing about your own experiences making the transition, or similar programs you might know about. We'd also be interested in hearing about what, in your experience, are the most important programming skills needed by someone with a good knowledge base but little practical experience to be a strong contributor at their first job or internship.
72 comments
[ 1.5 ms ] story [ 232 ms ] threadBeginner - Anyone with an interest similar HN. Maybe resource to get into the Learning Area.
Learning - Place to find others learning the material. Maybe find other people study with. or collaborate/reproduce projects.
Experts - people actively looking for employment(what you already have planned).
We're starting with 2 tiers instead of 3, but the goal is similar.
Or if it was a clever reference to something in AI that made you sound smart but because it was a clever reference comes off more funny than rude.
> To qualify for joining, you do an online deep learning quiz (here: https://www.sharpestminds.com/members/apply), followed by a technical interview.
> Once you join you get access to a job board with exclusive (i.e., not scraped) internship and full-time opportunities on it.
These three constraints don't reconcile with each other.
Yes, new ML/AI resources like TensorFlow and MOOCs have made AI more accessible, and that having a degree is no longer required to implement ML/AI. I agree it's unnecessary gatekeeping to require a degree to be able to play with ML/AI.
But what showy YouTube videos and Medium thought pieces don't teach is implementing ML/AI in practice to solve business problems. The stereotypical quiz + technical interview for the ability to join the service won't account for that.
When I was looking for jobs last year, 100% of the job openings for ML/AI (as opposed to Data Analyst/Data Scientist) required a Masters/PhD. In that case, I can't blame them, since there is a certain amount of experience and knowledge required to define problems and work up statistically sound solutions that can't be done by simply adding layers to a neural network or ensembling XGBoost models.
Having an MSc / PhD in the field doesn't resolve this. HR departments use grad degrees as first-pass filters, and thereby miss self-taught people who are genuinely competent.
We try to solve for this by easing people into jobs with internships and work terms first. The community is a key part of that since it supports them if they get stuck on an implementation problem. And of course we're incentivized to make sure members perform well in the internship phase, since we make money when they're hired full time.
(I'm guessing.)
In web dev, experimentation is cheap, so you make changes fast and see what happens. In many ML applications, trying stuff is expensive in time and/or money. So the best strategy is often to think hard about what could be going wrong, and make and test explicit hypotheses.
The difference definitely came as a surprise to me when I was making the transition.
For a long time as a non-web dev, I reckoned that I was smarter than 99% of the web devs. Then a web-dev friend of mine made his first few million $, and I reached some humbling conclusions.
The YouTube/Medium posts however advocate "Learn Machine Learning from scratch in 3 Months!" which is a problem.
In that it reminded me of the way that Basecamp hires - which is that in the "final rounds" they actually hire the candidates to do some small project that is actually needed at Basecamp - just another way of getting at the bottom of what a person can actually do rather than how they look on paper.
Wasn't aware that Basecamp did this too. Thanks for sharing!
My Firefox doesn't run JS though, so there's that.
But yeah, a different name would be better.
- a few of the questions were very good, and either spoke to key high level concepts, or were specific while being language agnostic. (e.g which one of these layers wouldn't you need, why wouldn't this type of classifier work on this data).
- too many of the questions were hyper-focused on the minutiae of word embeddings, tensor flow syntax, SQL queries, and recommender schemes.
- many of the questions were constructed vaguely enough that "I don't know" would be the technically correct answer even though I don't think that was what you were going for.
metadata: recent PhD with serious grad courses in ML and working in DL/CV for the past year using a non-tensorflow framework (PyTorch).
We're constantly iterating on the quiz and it would be great to get more detailed thoughts on it.
If you'd like to do that, please get in touch! (Email in my profile)
High-level I don't think a quiz is necessarily the right tool either. Reminds me too much of taking the SAT or GRE.
Statistically, it does an OK job at being an initial filter. My biggest concern at the moment is that it's too coarse of a tool and it might be mistakenly turning away competent people.
Definitely a work in progress. If you have ideas on alternative formats or better questions, please email me. (Email in my profile.)
We've never had a work term work out to less than $25 / hour, except possibly in geographies like India where cost of living in USD is very low. This is below market, but it's only for the work term before getting hired.
Will update this with the most recent stats and clarify part-time status. Thanks again for catching this!
Yep, quiz might still be a bit too focused. Trying to strike a balance between breadth / depth / time spent answering questions, but still some fine tuning to do
Why does Facebook have to know what I am doing jobwise?
You can either (1) keep going with your current job and do this on the side assuming your employment contract allows it, or (2) apply to a company that wants to do full time right away without a work term.
There may be other options depending on your situation, so email me if you want to discuss further. Email address in my profile.
1. At least in the title of this post it is called an "Online Community" but it really feels like a job board. Is it really a community at all? I expect to feel disappointed.
2. Wait... is it a job board, or an internship matching system?
3. I got about 5 duplicate questions.
4. The timed questions with a big code block and multiple choice were stressful, in that there was some dense code to read and I couldn't decide whether to understand the code first or read the questions first.
5. It wasn't clear to me that the timer was actually a limit, and not just a suggestion (i.e., something to pace yourself to do all the questions in the time limit)
6. The SQL questions felt like very normal SQL questions. They seemed easy enough (assuming I got them right!) simply given past experience with database driven websites.
1 & 2. It's.. both. From the inside it feels like Slack + job board + GitHub-like profiles
3. Investigating this issue now, thanks for flagging
4-6. Noted, we'll keep updating the UI & question bank. Knowing this about how the timer feels is especially useful
Why only 'Deep Learning'? When you call yourself an 'Online Community for AI Devs', you need to consider projects from any domain in AI.
Deep Learning is just a sub-part of ML, to be precise, neural nets.
Please stop throwing keywords