Hacker News Reader

Comment by DiggyJohnson | Hacker News Reader

I feel like if you’re competent but you’ve been unemployed for a year now it’s because you’re unwilling to compromise. Firms are still hiring.

View on HN

No im genuinely having a difficult time getting any responses at all.

I've had three interviews since march. Method of application doesnt seem to matter, LinkedIn or Company Site, etc.

I am not entry level. I have seven years of professional programming experience. Ive written raytracers, a handful of scheme interpreters, autograd and ml frameworks, hardware test frameworks, lots of CRUD stuff professionally. Im language agnostic at this point, can do cpp, rust, python, julia, common lisp. Im an expert in Deep Reinforcement Learning, implemented AC, A2C, TD3, DQ, DDQ... APE-X, Distributed Batch PPO, have read over a hundred ml papers.

heres my linked in. https://www.linkedin.com/in/gibsonmartin

No responses. instant rejection from 99% of applications. interviews go well but i get ghosted. Unemployed since last october

not trying to be a dick, but after seven years you must have some friends or former colleagues that work at other companies, right? if so, why are you applying through the jobs page instead of having them refer you internally?

fwiw, I did take a look at your linkedin page. imo, you are listing way too many skills, languages, and tools. I don't doubt that you worked with all of those things. but when you list so many, it raises doubt about which ones you're truly proficient with.

I have only worked for small companies. When you work for small companies you have to wear hats. When I worked at skylark I had to write cli tools to interact with their SoC in c and bash, but also do hardware test in python, but also make a yield assessment website (js/python/orm shit). At Cratus i wrote a microservices backend docker yaml gui with a react frontend with a team of ten-ish turkish guys, but also worked with their electrical engineer on some embedded flashing webserver, a fluidflow calculator, and some ml camera tech involving nvidia jetsons for a taiwanese company, and a bottle manufacturing company for flaw detection. I was tasked with reverse engineering some chinese blue tooth beacons, and implementing a multilateration algorithm. I personally went to the Olympus endoscope repair facility, interviewed the staff, mapped the facility, and wrote a simulator for it.

I have taken the few jobs I have been able to get. I learn what is professionally required. It might sound crazy but this is genuinely what i worked on. Atleast I can say it has not been boring!

I've worked at small companies too, I get it. the charitable read of your linkedin page is that you're a engineer with a broad skillset who can be dropped into any kind of project and get it done. the uncharitable read is that you haven't spent enough time with any specific tool/language/domain to truly master it. you're not appealing to recruiters who are looking for X+ years of specific skill, and you're not telling your story the right way for recruiters that are looking for a generalist.

if I were you, I would actually remove most of the individual projects from each job. keep 2-3 examples of what you consider your most impressive work and sell the reader on why it was impressive. key points to hit are a) why was the project important b) how did you solve the problem, and c) what was the impact to the business. by the way, b) should not just be a description of the tech stack you used. if you're looking at senior roles, people/organizational problems are more interesting to the reader.

That is reasonable advice. Thank you.

I had it reduced to all back end development for a few months. I got no responses. To be honest I've given up on it.

As for my friend developers. One works at apple, one works at amazon, those are unlikely hires for me.

One got laid off this year and now is getting certified to drive a city bus. A few of them work for the BP subcontractor, which is currently not hiring, layoffs. My other friend just got a new job, he was unemployed for about six months this year. My other friend has a degree and years of experience but is unemployed.

I have one friend in bed with C# and medical industry. Have asked him about it.

Lets say the average person has 10 professional acquaintances, id wager many of them are simply not in a position to make a referal.

that's too bad. friends that can directly vouch for your skills are the easiest way into a job. still, don't forget a lot of companies give a bonus for any referral that results in a hire. unless you've specifically pissed them off, I bet most people who even vaguely recall that you've worked together would be willing to refer you. I got my current job through a second-degree connection. I didn't even know the guy that referred me.

My unsolicited advice

You come across as very technical from your LI. Perhaps the companies you are applying for are looking for software engineers that will make the company more money than their salary. convince me that your technical ability will allow me to create more trinkets at a lower price and higher quality (or whatever the company does) than my competitor and I would hire you in an instant.

I launched a rocket in factorio.

In seriousness, I have hardware test and yield assessment in my background. Ive been to multiple pcb lines, talked to techs, worked with electrical engineers on pointing out repeated failures in tombstoned resisters, burned out parts, etc... I have personally ideated, developed, and deployed tests which made higher quality trinkets for less money at higher success rates. Me and the Electrical Engineer at Skylark made the yield go from 30% to 99% over a few revisions. (Mostly because hes very good at his job. as the test engineer all i could do was point out the failures and collect data.)

My big project at Cratus was real time asset tracking for hundreds of thousands of endoscopes, with alerts on slow, lost, unshipped assets. Reducing turnaround time for hospitals all over the country.

I suspect that making more trinkets for lower cost is fairly well demonstrated by my resume and work history, probably more so than the average web developer.

Likely... hiring staff cannot tell.

I’m going to do you a favor here and be more honest than most will, please understand it’s coming from a place of good intention and trying to help you get a job: your LinkedIn reads like a crock of sh*t.

- Your tag line claims you’re an ML expert but you say you have 7 years of work experience and your LinkedIn doesn’t have any educational experience. I disagree with the necessity of degrees but 7 years of largely several month stints and no degree is hardly enough experience to be considered an expert in anything. If you’re into ML, that’s fine to put there but it’ll be off putting if there’s a stolen valor element to your profile.

- You have way too much written for some of your experiences. You were at CRATUSTECH for 7 months. If what you did was amazing enough to warrant what you wrote there, your previous managers would be knocking at your door right now. Pair things down so you don’t come off as inflating your contributions.

- You’re presenting your experiences in a confusing way. You have so much overlap that it’s hard to ascertain what your career narrative was. Most people who are experts in something weren’t able to reach expert level knowledge while jugging multiple jobs. Focus on your narrative so employers can determine if you’re the right fit.

- You’re commenting too much on stuff. No one is ever impressed by people’s comments online and the more opinions you share in comments the more liability it is for employers who already get to be picky right now. This isn’t to say you said anything wrong but if it were me, I’d just delete them to now have a busy activity section.

- Lastly, you need a better profile picture. If you want people to take you seriously put a better foot forward here. Even just a decently lit selfie. Don’t crop some photo of you and someone else out on a hike or whatever. Doing an activity is fine, but pose, smile, be by yourself.

Hope this helps.

-I dropped out of university because i wasn't allowed to take only compsci and math. 80% of my first two semesters were pointless state required classes so i left. That was 11 years ago. I love programming and I do it first thing in the morning every day for many years now. Deep Reinforcement Learning is a a very approachable field! With less than a year of full time effort somebody that has been coding for years can become an expert. There is no stolen valor here, I can read ML papers and implement their architectures from scratch. I have assisted with undergrad and masters papers. Have met RL experts who consider me to be a peer. They have been very friendly, and mostly just happy to find someone else interested and knowledgeable in the field. My DRL knowledge does not come from work. I have not really found an opportunity to use it professionally. I learned in a cabin in the woods. As for how to effectively communicate that I have this skill in a believable fashion despite not resembling a PHD I do not know.

- That is largely due to other people revising my resume. I admit I don't like the tone. It was very dry and literal before, but a lot of people made me rewrite it in this value-added tone that seems popular nowadays. I did not get employment attention when it was dry before, or with any of the other variants I have tried. This was me succumbing to lots of advice.

- I do not have a career narrative. I struggle to get a jobs and take what I can get, usually small companies looking for a deal on me.

- My wife took that picture. We got married this year. I could change the picture.

It gives me an example of how I am being perceived, but I don't think it will help. I don't think most companies are even reading my resume or linkedin. I suspect my application is getting instantly thrown away.

Have you asked those RL experts to offer you a job or connect you with someone that is hiring?

Yes of course I have. I always tell them im looking for work lol. They just arent in a position to hire me. I had a meeting four days ago with one, who has a small startup, but very low funds. He said he will check back with me in about a month, theyre doing fundraising now.

Seems to me like nobody has money. A couple of these small companies are telling me theyre hiring over fiver. Remote devs in the Philipines. I've been able to scrounge a few hours of freelance from them here and there, and some ludicrous offers of high equity, and basically no pay.

The advice from others is fine, but I think you're struggling because you only made it a year, 7 months, 11 months, and 7 months in your last four roles. Wherever you land next, you need to knuckle down and show some staying power or it's only going to get harder.

It has been circumstance. When you work for small companies there are sometimes problems. At the BP subcontractor there was the OPEC collusion that year. I saw an entire floor of Westlake campus get fired. I was a contractor, and I started to get questions about my bill-ability. My next position saw a very large bill come in one day, from a bad rev of an older board. We had a party when I got laid off. My next job I was the last person remaining in the office during corona. All the other workers were in Turkey. I was severely underpaid. The boss harassed me and threatened me, forcing me to work overnight multiple days in a row, and the weekend. I did leave that job, and rightly so. He continued calling me for many days for over a month.

The last 2 positions I had were for crypto companies. Those companies no longer exist.

I am 29 years old. My career started really in 2016 or so. Stable corporate America simply hasn't existed for me. I could not have kept my jobs even if I wanted to.

The explanation doesn't really matter at this point. Recruiters will filter you out at bigger companies, and smaller companies can't afford to take the risk. But sooner or later you'll get some chances, and I'm advising that you optimize for longevity when those chances come. To get out of this hole you need to stay put for at least two years, ideally longer.

All the other advice about how to represent yourself is good and will help you find that next role. My advice is for after you get it.

I think I was getting filtered out before regardless. Probably by machine.

I think it’s all just contract work.

My partner was a contractor but on their resume all you can see is something like “9 years Booz Allen” because they only credit the agency as the employer.

To the OP, if you has an agency then credit them and remove all your gigs that are less than 1 year.

Also if you don’t have a degree, employers subtract 4 years from your experience. College isn’t totally useless.

Lastly zero people will believe you are a ML expert without having a long background in a company or having authored a paper. ML isn’t web dev, it’s something that a more of an academic cutting edge now than what just anyone can pick up.

A lot of it was contract work. Some of it I didnt even have a contract. I have no agency. If i removed my gigs less than a year then my resume is going to be almost blank.

Secondly, on the ML front. I didnt just pick it up. Firstly, I practiced on and off for years. I failed to learn it handfuls of times, trudging through formal education resources about integrals and backprop and matrix math.

Eventually on one of my unemployment stints I found this book called Grokking Deep Learning and drove across the country and went into the redwoods in california, and I studied it for months on end. No internet on my desktop, only my laptop for looking stuff up. With no distractions the amount I accomplished was great.

Forgive me for getting angry but Im starting to get mad that people don't believe I could exist. I woke up at the god damn crack of dawn every day and studied ML, implemented autograd, my own little pytorch, and every architecture I could find, and read a zillion ml papers. And when I got bored in the afternoon I went and hiked up the mountain. Im a real human being. I am an RL expert. I dont think im special. I believe anyone else can do this too. People just don't know what they are capable of. They won't believe me because they don't even believe in themselves.

I don’t think you don’t exist. I don’t think you didn’t work hard for your skills.

I just don’t also think you are an ML expert.

Yes, you know what you are doing and have some learning projects under your belt but so do university students. They even write formal papers and publish them under the guidance of people who lead the field. Yet they wouldn’t claim they are experts!

I have had handfuls of university students come to me for help in the full pocess of deciding what to make a paper about, coding the architectures for the paper, and reviewing the paper. That includes masters papers,and it has sorely affected my opinion of formal educations ability to teach DRL and its reputation.

Most Ml, and especially DRL papers are useless, or worse, lies. I would say over 60% of them. The authors are usually motivated by graduation, or social situation, and often contain faked results. This is especially true of deep RL papers. Deep Reinforcement Learning is hard.

You cant just fake the data of the model succeeding because the paper is due (Saw this three times, fought them). Why are you using a discrete action space for a continuous environment?? (Usually its because they cant get the code to work for non Q based algorithms. Until PPO came out the entire Actor Critic family was really hard to converge, and everyone was failing with all the variants of DDPG. After PPO that changed, but the average ml paper writer couldnt succesfully implement it. I failed a few times, took me two solid weeks of hacking at it. Students who fail to implement their AC family algorithm would go use DQ from the 2012 Mnih. paper because they learned Q learning on gridworlds in class, but it only has discrete outputs... not only that but DQ was replaced by DDQ in 2015, and under no circumstances other than education should the original DQ be used. Real DRL just has almost nothing in common with the gridworld and Tabular Q based classes people are taking in school.) Why are you truncating the model outputs instead of letting the channel between two models be in feature space, despite not using those truncated outputs as some sort of gradient constraint? (Many companies are still doing this) Where is the epsilon decay schedule in your paper?? and basic entropy graphs, convergence graphs... you cant just publish the env score. you should ensure convergence on a control environment first. No you cant continuously shape the rewards manually throughout the paper to raise performance. No your model wont converge when your environment has sparse rewards or prior gates. No you cant start your discount factor at zero. (lol)

Some of the university students I assisted were absolutely horrible at programming, gave up early and lied in the results. Now I dont have a problem with people practicing and learning but it was clear to me that this wasn't the right setting and they didnt have the right resources or oversight to succeed, and that sucks.

But people see their paper anyways, assume its valid, and they got a degree that year, so I guess I better bow to them?

Whats your criteria for expert? Being in the top 10% of DRL practitioners is a fair measure. I was certainly in that fraction at the peak of my study. Relative to 99.99% of the world population, and 90% of the people who even know what DRL is, I am an expert. I am an expert in Deep Reinforcement Learning. If you have any paid projects you need help with let me know.

Reply on news.ycombinator.com