My Financial Story – Chapter 4 – Dream Job Number 1: Research in Robotics

This is the fourth chapter in my financial story series. In the previous chapter I became an engineer and engineered my first financial failure. Here I’ll show you how dream jobs are not that dreamy. In next chapter we’ll destroy another dream.

Note: I have stolen copied been inspired by “the job experience” post series by livingafi. I loved reading his series and I’m trying to do something similar here, where I analyse both my work and finance history/goals.

So far in this series I’ve been restricting the focus on the financial aspects of my young life. From now on, the focus will be on both finances and jobs. This blog is about money, work, passions and happiness as I see them: reaching Financial Independence in order to Retire Early and follow your passions. (Un)Happiness at work is one of the key ingredients here, so I think going deep with my work experiences can be both fun and insightful.

I’m going to call this sub-series Dream Jobs because hey, guys, I’m a lucky man and I’ll tell you why: a lot of my dreams became true. Problem is that they happened to be waaay less dreamy than expected. What was I dreaming about? Well, any decent dream of mine revolves around doing something to make the world a better place, having big impact following one of my passions. I’m not that kind of guy who dreams a mansion or a Ferrari. I dream about a better life, an ideal life, for me and my fellow earthlings.

robocopAs a kid, my dreams were guided by role models. Factory of which, before the great venue of the Internet, has been the Cinema. One of the movies that impressed me the most, that I watched two-digits times, was Robocop. I loved robots. I spent my youth staring at Goldrake, Mazinger, Steel Jeeg and many others. I grew up thinking/hoping that one day, finally, robots will be among us for good.

I reached age 25 and they were not around yet. Where are they? There should be a Fermi Paradox for robots, shouldn’t it? So when it was time to choose my Master Degree curriculum I picked the most AI / Robotics oriented options I had in Software Engineering.

I had no idea what to do after my master degree, so why not try to follow my dreams? In spring 2002 I met my future supervisor and it was “love” at first sight: “our research group attends robotic competitions… Robocups… Autonomous robotic… exploration… mapping of an unknown environment… there’s this conference in Acapulco next year… we are hiring…”. Acapulco?? I had to look it up on a map where it is! I cannot not apply for a position that’ll bring me to Acapulco to talk about robots and Caipirinhas. So I ended up starting my academic career. Just, as a good Italian does, following my gut feelings instead of my brain.

At first, the research group seemed exciting: there were young and fun researchers to work with, with whom I quickly connected. So I decided to join the lab in summer 2002. I was still missing 3-4 exams before my master thesis but who cares? It took more than expected to complete my curriculum but in the meantime I started working on actual robots! I picked SLAM problem (Simultaneous Localization and Mapping) as my Master Thesis topic. I lectured a couple of courses – computer vision and reinforcement learning – while I was still a student and it was fun. I started seeing a pattern here: I like to interact with students, I like to teach, I like to share. And apparently students like me too.

pioneerBeing myself a student in our research group was awesome: no responsibilities, robots availability, freedom to experiment, freedom to code. Heaven! Well… not really.  When I joined the team was deeply committed with building their own robot, so there were several mechanical and electronic engineers in the lab. They designed a super-ultra-uber-powerful robot they were not able to build and so they never finished. This project lasted till my graduation in September 2003 when finally the supervisor fired them all and we purchased a true mobile robot.

padovaFun Story: well, actually it’s a very sad story. It involves my master thesis and my first Robocup competition in Padova (Italy), summer 2003. I had completed my master thesis and eager to experiment my algorithm on our work-in-progress robot. The robot never moved more than a meter in the lab in last 9 months. I wrote a simulator for the robot as sandbox to test my code, but I wanted to see the code in action on the real robot. We were not sure it was going to work. In crunch time, just before the Robocup, we (actually, they) worked 3 consecutive days and nights in the lab. The very last day, after few blackmailing yells by the supervisor, she decided we’d go to the Robocup and fix the robot there. We rented a van and reached the luxury hotel, with swimming pool and spa, just to drop off the luggages and then headed directly to the competition area. Another 3 frustrating days trying to fix the robot, when everything was clearly broken: a robot designed to weight 20kg that ended up slightly below 100. Actuators that were undersized. Batteries that were too heavy. Sensors that were damaged. No way. acapulco-beachesWe never took part of the competition, we slept on the paddock area, working 24h a day. Well, they were doing this, the mechs&electronics. I was just waiting, waiting, waiting,… then the competition was over and we went back to our hotel to grab unused luggages, and headed back home. I’ll never forget the sight of the swimming pool, my smell, my sleeplessness in the same decadent picture. The day after, in the lab, the supervisor fired all the m&e engineers and blamed everyone in the room. Bye bye Acapulco. Fuckapulco.

The supervisor claimed that the last year was a waste of time and her own money. We will never do it again. Let’s buy real robot and let’s focus on research instead. I bought it and after my graduation (end of September) I tried to join the faculty as PhD student (mid October).

To become a PhD student in Italy you have to pass tests. There are fixed number of open positions, some of them with a faculty granted salary, others without, and you need to study a lot for it. I had no time, just 20 days between final graduation and PhD test, with a couple of necessary graduation parties in between. I still remember people at my graduation party saying “hey, congrats! You’re now an Engineer!! Now you can finally take a break and relax a little, you have finished!” No, I didn’t. “Well, actually I have the PhD tests in 15 days and I plan on studying for it something like 12 hours per day so no, I haven’t finished at all…”. That year was the last one of free lunch in Italian’s academic world. ~50 applicant for few positions. No, I didn’t make it. I could and should have relaxed instead.

I accepted a 6 months contract (1000 Euro per month) with the research group with the goal of finding other funds. Didn’t know in what kind of hell I had thrown myself. In these 6 months I realized that the supervisor was a total asshole. Manipulative, aggressive, offensive, incompetent, workaholic. A classical example of “Italian meritocracy” in action. Even within Italian general incompetency-based equilibrium she was the worst. The black sheep of the department. But she was rich, very rich (thanks to her family), so she afforded expenses with her own money. Robots? She bought it. Conferences? She paid the tickets. Lab? She rented a place. We were at her own crazy will. You may not see her around for few days but then, suddenly (usually on Friday or before your planned out-of-office time) she drops from nowhere and destroys your day-week-month. She was like the “hurry up” skull character of Bubble Bubble.

hurrySix months passed slowly. We were applying for every conference, every journal, every call for paper at the very last minute. Every other day we had to activate the panic mode.

HurryUp: “Urgent, super urgent! Need to submit something for conference XYZW by tomorrow midnight CEST. Send me 5 pages about solving problem X with methodology Y ASAP“.

I (or whoever was on target): “Well, 5 pages are a lot… and this is the first time I read about methodology Y, we actually never ever solved problem X with Y so what am I supposed to write? We don’t have any data, we’ve done no experiments, leaving aside that I don’t even know what Y is. Btw, we are not even solving problem X, aren’t we? You said last week that ‘it’s unsolvable‘, and the week before that ‘it has already been solved and it is now a non-problem’. Anyhow, facing problem X requires having a f*cky robot that actually does something! Ours doesn’t even move because we need time to focus on writing the actual code instead of jumping on every visibility/funding train (and missing ALL of them).

So I spent my time in the lab switching between:

  • Writing papers about things we didn’t do.
  • Writing proposal for partnerships with companies promising them things we weren’t able to do.
    • One of these actually accepted and we ended up working way more than planned – because we over-promised – and getting paid 30% of what agreed. 18 months, 3 people (part time), several on site visits (25 kilometers away) for 4800 Euro total. Gross. I’m not joking…
  • Reading other research groups’ papers and watching their videos to feel both inspired and frustrated.
  • Hacking my supervisor’s Matlab code into something usable, making her believe that her nested-10-times for loop was performance ok. Fun Story on object recognition code. Hurryup: “hey, guys, look at this! I’ve solved skin recognition with texture pattern matching and Gaussian Markov filtering bla bla bla! Look at this: here’s the test set of 10 images… yes, 32×32 pixel, I don’t know why with bigger images it crashes (MrRIP: really? You solved skin recognition? 32×32? Yes, I’m sure 1024 pixels carry enough information! And p.s. you wanna know why it crashes with bigger images? ‘Cause your code’s complexity is O(n100)!! ), but look at the results… here I check the code against image1… here you see a visible arm… ok.. let’s wait… I don’t know why it’s so slow… wait for it… why it’s so slow… maybe the wifi in the lab is not good… here it is: SKIN RECOGNIZED! Success! It works with all the 10 test images!!” blackMrRIP: “Wait, every image tested so far contains some skin, maybe too much. Why not try with an image that doesn’t contain skin and expect a negative outcome? And, btw, before claiming we solved the problem maybe we should try with some complex picture. I don’t know, something like ‘darker skin’ or ‘less visible body parts’,…” Hurryup: “why should we try to detect skin where there’s no skin? This is stupid! I thought you were smart…” texturesMrRIP: “yeah, ok, I’m probably dumb. But let’s try your code against a texture 32×32 monochromatic black… SKIN RECOGNIZED… LOL let’s now try a texture of a brick wall… SKIN RECOGNIZED… see dear genius? your code is a very slow function that just {return true;}!“.
  • Writing low level c++ code to make all the robotic hardware work. This was the only part that I actually liked! I wrote C++ code for SLAM, code for robotic Navigation, code for connecting the hardware parts like cameras, sensors, actuators,… that was the only reason why I lasted so long into the research group. It made me a strong software engineer and planted the seeds for my career. It was the technical part, not the research.

After 6 months I burnt out and rage quit. Hurryup offered me a 2 years contract but I said no, thanks.

I took a month off of work related thoughts, April 2004. I tried hard to detach and clear my mind but I had no knowledge of how the world works. Everybody around me asked “ok, cool, now what are you going to do?” and I was scared. I feared that I shouldn’t wait too long before jumping back into the supposedly always-accelerating train of my career. I did an interview for a crappy company that offered 18k Euro gross per year, a shitty job, and I rejected the offer. I didn’t know what to do.

At the end of my sabbatical month, Hurryup called me saying my contract was still waiting for me. We can discuss the issues I had and we can try to fix them. It was a lie, but I was so confused and unsure of my value. Plus, she added extra bonus salary (700 per month), paid out by her directly, for the first 6 months at least. I was going to earn close to 2k net per month. I accepted. In May 2004 I was back in the lab, confused, part of me hating myself, part of me surviving in quiet desperation.

libsonaBut things became fun for a while. We focused on 2004 Robocup in Lisbon where we got close to win the Robotic Rescue competition for autonomous robots, our all time best score. The robot was solid and I wrote a lot of code for planning, navigation, localization and mapping. We, researchers and students – not hurryup, became a team of friends more than coworkers. I created connections that are still intense nowadays. One of these friendships will help me escape my academic career. Another stimulated me on a daily basis on code design, code quality and advanced C++ techniques. I read a lot of books, wrote tons of lines of code and became a very solid software engineer, able to win higher level interviews. I owe him good part of my fortune, thanks A.C.!

October 2004, everything was relatively fine and I was having enough fun at work. I was lecturing a couple of courses and working on our robots (yes, plural, we – she – bought a new one!). Not much research activity, but it reflected my desires: I’m not a researcher, I’m a software engineer. I love to write code and to see things work. I’m ok with being a research assistant and not a pure researcher. You write crappy Matlab code pretending it’s revolutionary, I write solid C++ code that handle 99% of the robot activities. But that was not part of the evil plan of Mrs. Hurryup.

She came into the lab, and wanted to discuss with me. “RIP, it’s time for another attempt at PhD acceptance test. Let’s discuss the plan. This time you’ll make it, probably without salary – first 6 spots are already allocated – but you’ll resign from the current contract and we’ll make a PhD salary out of that.

Too many things were wrong with that sentence. First, I essentially had to give up 1200 Euro per month (base salary for a research assistant) to give myself 860 Euro per month (salary of a PhD student in Italy at that time). Second, how do you know, exactly, that first 6 salaried spots are already allocated? Isn’t it a public competition? Well, turns out that there’s an internal black market among professors where research groups used to pay with their resources to have their students get the salaried access to PhD. A kind of Mafia, yes. And Hurryup, the black sheep of the faculty, never had access to it. Third, I’m not sure I want to start something more research oriented that lasts for 3 years at least…

But it was not a question. I had to try. And I did. There were just ~20 participants, way less with respect the year before. In reaction to 2001 crisis the government started killing resources and positions in the academic world. It was not appealing to students anymore. There were 6 salaried position plus 4 nonsalaried. I ranked 7thI was the first of the nonsalaried. (Mark Watney‘s voice here) surprise

But then something unexpected happened: one of the “top 6” quits before starting. I received a letter saying that I won the salaried position. Before I could celebrate I received a call from Hurryup: “RIP, listen… professor HickDead, supposed supervisor of the guy who quit asked me to ask you to decline the salaried position and take the nonsalaried one. She funded the position and she wants the money back

Waaaat? These are black market issues and I don’t give a sheet. I participated a public competition where it was stated nowhere that the prizes were already allocated (it would have been illegal). I won it. I received the prize and you want me to torn it down?? No sir, deal with it.

Anyway I had to resign as research assistant in February and Hurryup stopped giving me extra money. I switched from close to 2000 per month to less than 900. And the hell begun. Being a PhD student meant that I was her property. I lost freedom of choices. I became her puppet and my life drowned. My research topic changed 5 times in the first 2 years. SLAM. No, let’s do Machine Learning. No, none cares about that (what an idiot), let’s do computer vision. No, reinforcement learning. No, pattern recognition.

robocup2005I woke up most of the morning asking myself where I was heading to. No more working on the robots. We were one of the strongest teams in 2004 Robocup. Rescue robotics was attracting money and attention, both Japan and US were investing money on it. Robocup 2005, Osaka, would have been the robotic event of the year. We could and should have been working for it. We didn’t. But yes, we went there and we sucked. Well, they went there. It was too expensive and Hurryup paid for only 3 team members and I was not one of them. I never traveled outside Europe and I was one of the most expert tech members of the team. It was one of the saddest moments in my life. In the mid of a competition they had problems that only I was able to solve and they called me. In the mid of the night. I don’t remember what I replied, but something like “f*ck you”.

Things went south rapidly and passion for work declined like the stock market in the great depression. Quitting the PhD seemed the only wise thing to do, but anyone, literally ANYONE, kept telling me that it would have been the end of the world. I may even be sued to repay the faculty with the salary I cashed so far! My career would be ruined, my resumé will be worthless, etc. I felt doomed and kept going on.

bremenSomething good happened in spring 2006. We had a bunch of students joining the lab and I helped few of them with their master thesis. The same pattern again: teaching, helping others, giving. I’ve always been good at it because I loved it. The newcomers wanted to play with the robots and had time and resources to actually help. Last Robocup we attended as a team was Bremen 2006. Bremen was relatively close and cheap so we all went to the competition. Something like 15 of us. We made ad hoc T-shirts with our names printed on. I felt part of a team again. Obviously, robotic research advanced substantially in last few years. We were not competitive anymore. German, Korean, Japanese, American teams with sponsors behind and serious budgets were leading innovations. We passed first round but were kicked out as soon as we faced decent teams.

Most of the students disappeared after the Robocup or after graduation, by end of 2016. Hurryup claimed we – meaning us, not including her – were not good enough with robots and maybe we should shut the robotic lab down to only care about computer vision. I was at the beginning of the third and last year of my PhD with almost zero relevant publications and research topic just changed for the 10th time. No way I was going to complete it by end of 2007. Salary will end in September and there were no projects around to pay me. Hurryup knew I was locked in and had to keep going, even without a salary. It would have been too stupid to quit now… Wouldn’t it? I was close to age 30 with a salary of 860 Euro and months of free work in the near future, with almost certainty I won’t get a job in the academic world (at least the Italian one) due to the upcoming public job crisis that still lasts today. The future looked so dark...

I spent the beginning of 2007 hating myself, hating Hurryup, hating robotics, hating academy, hating my accomplished dream. No, it had to stop.

marioA friend/colleague of mine that we’ll name Charon – one of the big waves of 2006 – moved to Milan just after graduation to follow one of his dream, that was one of mine too. We used to play computer games together. First person shooters, real time strategic games, role playing games,… we spent a lot of time together just playing and daydreaming about becoming videogame developers.

He did it. In March 2007 he was hired by one of the biggest Italian videogame companies. I was happy for him but felt alone and lost. After one of the deepest internal crisis ever I decided I’d try. I went to Milan for an interview and I got the job. They offered me 22k Euro per year, I asked 26k and we reached an agreement at 24.5k. Circa 1500 net per month, starting in July 2007 to follow the next dream in my agenda. A dream to be explored (and destroyed) in next chapter.

I didn’t quit soon the PhD. I asked Hurryup if I can get an extra year, obviously without salary, and I’ll focus on a topic I feel passionate about: Artificial Intelligence in videogames. I’d work in the evenings or in the nights if necessary. I wanted to develop AI for Turn Based Strategy games, like Battle for Wesnoth, an open source TBS (Turn Based Strategy) game I still play and love. I sent her a mail filled with enthusiasm, passion and energy. She replied with a mail full of hate and misery. She threatened me with the salary story “we may sue you and you’ll have to pay your cashed salaries back to the faculty!“. Lies, as usual. Goodbye Asshole. Fuck robots. Fuck society. CCed on this mail thread with me and Hurryup was my PhD thesis reviewer, that for now on I’ll call him Brofessor. He replied personally to me: “hey RIP, I saw you’re passionate about videogames and AI… we are planning to launch a Master in Videogame Development in the near future in our University. Would you want to be part of it?“.

The night was dark and full of terrors, but things were turning on my side, finally!

From a financial point of view during these 4 years I continued to spend less than I earned, even though my earnings were very low for most of the time. I didn’t keep fully active the private lessons machine: it’s a job on its own. Students grow old and you need new students coming in. I didn’t invest much time on it, so I had few students. I didn’t do any major financial mistake, I didn’t buy a new car, I simply kept half of my wealth all cash in a checking account and the other half invested in low interest CDs.

First year was super, due to 1200 per month salary and extra 700 per month Hurryup bonus. Then PhD hell came, making my economic life mostly flat.

November 2003 – July 2007

I don’t have recorded data before November 2003, but the ramping up started around June. You can see a drop in April 2004, the sabbatical month, and a nice climb up in May-Dec 2004, before the PhD hell.

Summarizing everything up to 2007 you get the following:

finoa2007
1991 – 2007. You can still see the impact of 6.000.000 Lira’s for my 18th Birthday and the bad Investments in Italian stocks in 2000-2002

I went to Milan on July 1st with 26000 Euro in savings, seeking for fortune.

Fortune we’ll see I’d found there. As always, thanks to people and not to the job.

4 comments

  1. Hi RIP,
    Just started reading your blog about a month ago. I was particularly curious about this chapter since I (too) fell for Academia and its shiny “aura” in our beloved Italy. Don’t need to give you any details, pretty sure you know them all.
    I finally decided to leave it in March (great timing) after three years as a PhD student and two years as a post-doc (and a five-month “continuiamo a lavorare al progetto in attesa che versino la quota dell’assegno” hiatus in between them). Got tired of waiting for “when the timing is right” and living month-by-month, wasting a ton of money in the process.
    Now I’ve been doing consulting for a few months and – although leaving the ivory tower looked terrifying in the beginning – it’s starting to work out pretty well, and I can say I’m starting my journey towards FI.
    Now I’m looking forward to read the rest of your story! Thanks a lot for all the content you’re sharing.

    F

    1. Oh brother, I know very well that “continuiamo a lavorare al progetto in attesa che versino la quota dell’assegno” feeling. I’ve had at least 2 colleagues in that situation. I quit before reaching that point 😉

Leave a Reply

Your email address will not be published. Required fields are marked *

Comment Spam Blocking by WP-SpamShield