I’m looking to get a crash course on a few topics and am hoping to do that by sitting down with domain experts for intensive 1:1 sessions. Has anyone here approached learning in a similar way, and if so, how did you go about sourcing experts? Currently looking for an AI expert.
I've had great luck with graduate students (and sometimes even ungrads): I'll find a grad student who knows the topic, often by searching for lesser known content creators on youtube/blogs, and reach out with an offer to pay for their time to explain a particular topic. So far it's worked out great, and I've created relationships with smart up and coming people in the field - they make meaningful money, I accelerate my learning. I ended up eventually hiring one, and another did consulting for me for some time. We had a routine where I'd ask them to read and summarize / teach me ML papers that were of interest to me, which they in turn could use in their studies/thesis/youtube channels.
Tips on this: content creators tend to be more open to and better at explaining things, and you get to see their ability to explain before you pay them. If you can, overpay them - students need the money more than you do :-)
This is very good advice. It works even better if the concept you are looking for is on the MS level rather than PhD candidate level because its pretty easy to beat T.A wages in most jurisdictions.
This is also a great way to do interviews if you have a small number of candidates: Just make them explain a final year BS / first year MS topic from the beginning without handwaving.
I wonder how you word your cold emails so that people bother to respond. Do you offer an hourly rate upfront in the first email? Thing is as someone who’s quite frequently on the receiving end of cold emails, I respond to companies looking for contracting/consulting because I know they probably have an at least passable rate in mind, whereas the few times I responded to individuals offering to pay for mentoring or feature development in my open source projects, after a few back and forth I realized they expected to pay one to two orders of magnitudes less than what I normally charge (e.g. $50 for about a week’s work, true story), it’s frankly embarrassing, so now I no longer respond to this sort of individual requests.
Two had small youtube channels delving into details of ML papers, one had a blog with a good explanation of a specific detail on a paper I was interested, one was a recommendation by my graduate advisor (I keep in touch with the university I graduated from a thousand years ago). The undergrad had a fantastic comment on a thread somewhere, I think in Kaggle.
I have an old boss that calls me from time to time. Maybe I'll just call him an old friend at this point. He lives by the philosophy that you put in the work and you debug your thinking by bouncing the context you've built up off of experts. Sometimes he shoots a cool hundred my way thereafter, sometimes he doesn't. Most of the time we're just catching up.
This works for him and tbh it works for me too. I guess my advice is that the important part is not sourcing the expert it's putting in the work to come with enough context to get something out of talking to an expert and to leave them without the feeling like you've wasted their time. Follow people on the socials, read their code, show up at NeurIPS with actual good questions to ask people in person on the hallway track.
Without the _hard work up front to get good questions to ask_ you're in danger of finding a good expert and them deciding you're just another starry eyed kid that doesn't know for just how many years longer you wouldn't even pass the screening call for an interview.
That sounds pretty good. Someone I know, calls me for a take, maybe offers a dinner, etc, sure. Even sounds like fun. A transactional 1:1 domain expert call, less so. I have done them from time to time but it's either been as a favor or prospecting for future business but isn't really a business model in either case.
ADDED: you've also described how networking actually works as opposed to the stereotypical networking event where a bunch of people desperate for jobs all show up which are basically useless.
https://Codementor.io is a platform designed for exactly that. I am registered as a mentor there, though not so active. I have not used it as a mentee, so cannot vouch for the process or mentor pool though.
AI is very broad, what exactly are you wishing to learn/build/accomplish and where are you in the process?
I specialize in ML that intersects with time-series, sensor/IoT data and audio. Info in profile if anyone is interested.
What have your experiences mentoring there been? I’ve been considering checking it out - do you have to do much to ‘sell yourself’ there in order to get clients? Do you have to deal much with poor spoken English skills?
A bit of a mix to be honest. The practical/technical aspects of the platform is good. There is some basic quality checking (and guidance) of mentors - good without being onerous. And the 1-on-1 experiences I have had were good. But I found it difficult to get a reasonable amount of gig the few weeks I tried it more actively. Majority of the requests posted are very low effort, barely two coherent sentences about what they needed. For a subset of more serious requests, that fit my skill set, I wrote personalized messages describing how I could help. I also made clear I offered half price for students and early professionals. But the hitrate was extremely low, maybe 2%. Probably my rate was too high for those requests. But it is already set lower than what I charge for regular longer term engagements. Maybe it looks bad I only have 3 reviews, idk... Overall I have found that it is not good as an added income source (as someone in a high CoL, Norway) - but it is great as a low-friction offer for those that find me and really want a bit of my advice in particular. They then find me via GitHub etc
I have been doing a few mentoring sessions there and it has been a great experience. However, I don't see that many people asking for mentorship there, I wish the volume of mentees was higher.
I did some mentorships there (as a mentor) a few years back.
My experience was "meh." Most of the mentees were last-minute students looking for help with their assignments.
But their SEO must be excellent. I still get web traffic and referrals from there.
It depends on how much money you have. You could pretty easily pay someone on Intro to spend some time with you, but experts can be crazy expensive. You could also just do a ton of cold email with very specific questions and build your own network of mentors. I'm sure posting stuff like this on various forums could also end with someone volunteering their time.
The more specific you can be with what you want to learn, the better off you'll be though. "AI expert" is still pretty broad.
"Life punishes the vague wish and rewards the specific ask"
yeah sadly capitalism is brutal about this. In the past mentoring and apprenticeship was based on the fact a business owner can get years of cheap labour out of you.
The economics dont really math out with software development, because to be a good software developer usually means you are smart enough to see when you are being exploited and nope out as soon as possible.
Then mentioning AI is an even deeper layer of "nobody wants you to know the tricks that is making them scam money"
Just "start doing it" and quickly you'll run into a wall, ask for help with a legitimate question that isn't vague like "teach me about x".
Solve that problem and move onto the next. Via a string of problem fixes you gain more domain knowledge and you'll retain more as you struggled through it and deeply understood it.
Heh, a couple of years ago, I had an idea for an "Uber for Experts." It would provide a similar experience to Uber, but instead of a ride, you'd get 30 minutes with a domain expert of your choosing. I never got around to working on it, but there might still be an opportunity for something like this.
That is more indicative of Google’s culture and lack of focus than the viability of the idea. With the death of Google Reader, plenty of RSS aggregators popped up and are still kicking.
There's a few companies out there that provide these "expert networks." GLG or AlphaSights
Not necessarily tutors, but I run into the same issue in Product Management. I need to do customer research, but the process of finding people to speak with is time consuming and very much like sales. To have 10 conversations, I'll probably have reached out to at least 40-50 people. I have to build a "funnel" of people in order to maintain having a few conversations per week.
Note: I experience this problem at startups that don't have existing customer bases. With companies that have existing customer bases, finding people to validate ideas and get feedback is not arduous .
I was an AlphaSights "expert" on cloud for a while, they paid me super well and the people asked good questions, I stopped doing it because It was mostly hedge fund managers I was talking with, but given how much they paid me, I can't imagine what their clients must have been paying.
I did a few calls with an expert network firm once upon a time when I was an analyst. My limited experience was that they were really looking for inside information rather than broader insights and a lot of product companies actually forbid engaging with them.
Super-well paid is a matter of perspective. As I recall it was about $500 for an hour call but it's not like I personally got all that and was pretty normal as a consulting rate.
Mine as after DigitalOcean went public, I was long gone from it, they mostly wanted to talk about where cloud is going (would managed hosting/ftp hosts go away, what do you think will happen to the front end, blah bah blah) - They paid me hourly well over an order of magnitude more than what you were paid, heh. :)
People could put up basic profiles with their skills listed and you could purchase a time slot with them. It’d be relatively low commitment for both the experts and also the advice seekers.
Some technical problems might be verifying expertise but this could be handled with a sort of social proof like how LinkedIn allows users to vouch for certain skills. In fact you could probably facilitate account creation by pulling from the LinkedIn api.
I agree. I’d like to be able to share my knowledge from time to time for a small fee and little hassle on my part.
I made some money on Codementor for a time and enjoyed it while I was between jobs, but wasn’t easy to balance that with FTE so I dropped it after a while.
Every now and then when I was working full-time, I had some piece work things that made me some reasonable pocket change. They were sometimes interesting (e.g. being a judge for a best-of award at a conference) and/or a way to fill a few dead hours and have an interesting chat. But if you have a good job, they're mostly a distraction for not that much money in the scheme of things.
Have you tried LinkedIn? From my anecdotal evidence, AI experts are the largest group there right now.
Irony aside, finding an expert for anything is difficult. Not because they don't exist, but because there's a lot of charlatans aiming to profit from the non-expert not really being able to tell. Do not underestimate their ability to BS even smart people, at least for a while.
My approach is usually to just ask around until I get a recommendation, a web of trust model if you will. I could for example recommend what I'd consider an AI expert to you. But you don't trust me because you don't know me, so by extension you have little reason to trust them. And that's what I'd advise: Find someone you trust, directly or by proxy.
If that's not a viable path, I would personally look around and try to figure out where relevant people hang out, write to one of them out of the blue. It appears people are more helpful to strangers than I previously assumed, though many do ghost. Many people have reached out to me this way, and I've sometimes ghosted them for embarrassing reasons like forgetting to answer, I've never once been annoyed by someone asking for my help.
I always offer payment, no matter how I found someome, based on the rate they name. I find that respectful, and some respectfully take it, others can't be bothered to write an invoice, and prefer to just do it pro bono. Though since you said it's intense, I would probably insist on paying them :)
- authors of leading textbooks or (open source or commercial) (hardware or software) systems;
- people that have many patents to their name in the respective area;
- scientists that have many peer-reviewed conference and journal publication to their name that are cited often;
- people that have or had O1 visa status ("outstanding scientist") or that served in committees for scientific bodies or governments, or as expert witnesses at trials.
I occasionally provide advice as a commercial service to find the right specialist for corporate clients for any given topic (knowledge brokerage).
I also offer a 1:1 "AI consultation" to CXOs in my own fields of expertise (machine learning, natural language processing, information retrieval, software engineering), pay per use or retainer when on the board of scientific advisors.
There is an ocean of tutorial videos out there, too, it all depends on how specific your need is.
Before talking to experts you could start out working through intro materials so you will have clueful and specific questions for the experts.
The old Python usenet group had an informal custom that the more knowledgable users weren't expected to spend much time on newbie questions. If you were a newbie, you'd mostly ask easy questions (say about the break statement) that intermediate level users could answer, so they would. After a while, you'd become intermediate level yourself, and then you'd get to answer newbie questions, while more advanced users would answer your now intermediate-level questions, and so on. Finally, the big time gurus were lurking about and they would mostly answer questions that really needed attention from experts.
I felt like I'd "arrived" when one day, Guido van Rossum himself answered one of my questions. It wasn't that great or advanced a question, but still.
So yeah, if you're a beginner, maybe you can just post your questions or watch some videos, or take a class.
As for beginner AI stuff, I liked the fast.ai videos when I started watching them a few years ago, though I didn't get through that many. They have since been redone, so if I wanted to get into the subject, I'd start with them again at the beginning. I don't think I'd try to engage experts until I'd gotten through all the videos and tried other ways to deal with any stumbling blocks.
You might have better luck at local events for likeminded professionals and network from there. Most people who are worth their salt won’t likely want to teach their craft to a rando on the internet. Nothing personal, there’s just too much to do.
It would help to know what your objectives are with learning more about AI. Otherwise we can only guess at your motivations.
I do this all the time, and the trick is finding someone who wants to share their knowledge. In a large company or govt this can be easy, as people have defined roles and expertise and you can find them and ask them.
Outside of that Upwork is generally where I go for any learning task where I need an expert for a short period. Good rates, 1:1 contracts and upwork overhead isn’t so high that it’s usury.
If it’s something super niche I can’t find there, then I’ll typically search for who is writing about the topic I’m interested in already, and then reach out if they have a public contact.
The last time I did that a few years ago I cold called a trash company in KS asking if there was anyone who would be willing to talk with me about getting into the business. The woman I called was the daughter of the owner and he invited me down to spend a few days in Russellville Arkansas learning how the trash business works.
I've found the key people at work, and took every chance to learn from them. But in general I've learned the most via blogs og people I identified as great sources (eg Joel Spolsky, Jeff Atwood, Bruce Schneider, and many others). I remember printing hundreds of pages a week, books and blog articles, to explore concepts and ideas (old days :p)
I'm on the other side now, and actually struggle to get visibility. I focus on Knowledge Management, not software development so much these days. I hoped to be in a new category on platforms like MentorCruise, but they considered it too nice. So I rely on my blog, my newsletter, and Gumroad: https://developassion.gumroad.com/l/pkm-coaching?layout=prof...
Really depends on the area of study, but in general it usually involves interning/co-op with someone you respect or serve in support roles.
In general, the wisdom of the tribe can't be given or taken... it is simply what you personally learn from the situation, and apply to internal problem solving processes. However, there are decent courses on specialty areas that often have a competent Prof/TA willing to chat with you about a specific subject.
Note, most "experts" are usually just perpetually improving students that see the bigger picture.
> Has anyone here approached learning in a similar way
Private tutoring is pretty common for learning languages, music, or for students in general. If you're willing to pay, it shouldn't be difficult to find a graduate students to teach you.
I recommend you get specific. "AI expert" could mean anything, from a visionary who funds startups but had never touched code to someone doing a PhD right now in computer vision.
Microsoft startup founders hub. I needed to get up to speed on RAG and Fine tuning. I shot about 40 introductory messages to their “Connect with an expert” network. I found a data scientist guy from Egypt who responded. He gave me a one hour crash course on RAG. I didn’t use his suggested methods but he set me up go and figure it out. I’m now caught up using haystack, chromaDB, DuckDB, and Mistral for my RAG needs.
So only applicable for students, but office hours are frequently extremely under-utilized, and professors love to talk about their subject of interest.
I'm not an AI expert, but feel free to ping me on other CS-related topics. See https://alexdowad.github.io/cv for a rundown of which topics I am 'expert' (or at least semi-expert) in.
Every time I have to do something complex related to home renovation, book keeping etc I wish there was a platform for this. Often you don't want to engage a firm for a week, you just need some advice, but you are happy to pay a healthy hourly fee for it.
I've had success learning the basics and not so basics of blender by going into the blender discord and saying I want tutoring on such and such topic and then working with whoever reaches out.
I've done similar. Check Upwork, codementor, clarity.fm, now also intro.co. Also post specific questions on niche forums and subreddits and DM the people who give the most helpful answers.
I've had luck on Reddit; usually programming related job boards or "who's hiring threads" - asking for hourly tutoring / pairing for specific things (e.g. rust).
I am an "AI expert" in the sense that I have been focused on applying generative AI for the last two years, (since we had useful general purpose LLMs).
Give me an idea of what you are trying to do and I will give you search terms to put in Perplexity or Claude or ChatGPT or whatever.
You are not going to find anything close to what you would get as far as value for mentoring as you would with LLM tools like I mentioned.
It wasn't a repost - the mods sometimes refresh a story to give it a second chance at more engagement. When you see the same post come back a day or two later, with the same comments, all with their timestamps updated to the current time, you can be confident this is what occurred. You can also tell by looking at the post history of OP - it shows an original posting date of 4 days ago, even though this post shows as a few hours ago. Not a bug, not someone spamming this site... this is a normal and fairly cool feature of HN.
This is a really bad advice. I'm trying to learn higher math, and sometimes I ask chatgpt for clarification. It's ok for definitions and explaining notation, but hilariously bad with proofs and logical reasoning. I have a collection of screenshots with chatgpt confidently spewing total nonsense (and sometimes doubling up or inventing something even worse when challenged). And this is for a topic where I can at least verify the answers.
Tips on this: content creators tend to be more open to and better at explaining things, and you get to see their ability to explain before you pay them. If you can, overpay them - students need the money more than you do :-)