My daily routine involves me going over a bunch of emails sent by different orgs and then chat messages and then going over the backlog of issues, PRs, etc. Part of the job is to reduce the backlog, make sure the product is operating as expected and customers do not have issues, another part is about pushing the new features and ideas forward. I am saying pushing because that is how it feels, it is quite different from the small tech problems where consensus is a matter of a few chats or calls. There are guardrails in place and people are positive but few have enough time to focus on an issue 100% and solve it in a day, instead tackling is spread into multiple scheduled events to align and make sure the proposals laid out in the documents make sense.
Some five years ago I was thriving in a small company and enjoyed myself building features and seeing them in production and customer hands quickly. The feedback was also quick and actionable. But I was getting bored and was dreaming about the grass being greener elsewhere. I am not regretting anything but just wondering if it is even possible to change your job satisfaction feelings when you know they should be good but the spark is getting smaller and weaker no matter what. Conversely I did not dream about going to Google or Facebook or any of these large corporate entities because I just knew it is not possible to move quickly and do amazing stuff there when surrounded by many many people, synchronisation should have to be quite expensive, and influence and politics play a much bigger role. My remuneration was also good so why even bother. But the bug was in me to “try it” and I could not shake it off.
2021 was a weird time to say the least. We had Covid lockdowns prior, the tech industry was booming after work from home had a huge demand on cloud services. It seemed as if everything changed and the future will never be the same yet everything was still similar and familiar. I got a job at Cisco and it was my first “big tech” experience, although some would argue the glory days for this company was over. I did even take a lower salary to jump there (think what Blind’ers would say about such a move). The company immediately struck me as being odd. Do not get me wrong the colleagues were great and well meaning. The infra work was also challenging and interesting, like global Kubernetes clusters, understanding how SIP works. But some old school java stack looked like it needed love. I tried improving it but when you need to present such ideas before 20+ people it becomes less of a fun refactoring but more of an attempt to persuade everyone it is a good idea when more or less everyone knows it already. Communication felt fragmented and sometimes it would be challenging psychologically to move forward and just do good meaningful and impactful work. Now that I think of it, I was probably in a shock after changing jobs. But at the time I found another gig at Microsoft, it was after less than a year in Cisco.
2022 was when I did a lot of interviews and got a place working in a team dealing with hardware identity management. Hiring manager was a strong motivator for me to join, he was great, although I did not apply to this position initially and had little to no knowledge of confidential compute or hardware attestations, they paid more as well. The part of the team was in Redmond (Washington) and the other in Ireland, we were remote. There were onboarding calls, training, and a bunch of reading to ingest the Microsoft culture and what exactly I was about to dive into.
Confidential compute is not as popular as it should be, I think. The idea to the keep memory encrypted and prevent the host from reading it is somewhat what you would want by default, especially when renting compute in the cloud. End-to-end encryption becomes your safety net to be able to deal with sensitive information without much fear of it leaking through the nuts and bolts that keep the cloud flying. Here comes the first main bright spot in large corporations, it is the ability to work on things which would rarely be in the firing line in a small outfit unless it specialises in this field.
I was happy to touch the reins and gallop with some of my new teammates after I got a proposal to lead the productionisation of an idea to capture the supply chain artifacts for the purpose of auditing. But to my surprise this internal gig got dissolved right when we were publishing internally the private preview version. Half of the team was gone, I got merged into confidential ledger, my manager got another assignment. This is another important lesson that big tech does terminate teams and you have to be quite vigilant to see it coming. You would see that in a small company miles ahead as the concept of runway cash is something you learn quickly. But in big tech it is not related to cash but rather to priorities and budgets at an org level which are quite opaque to a soldier on the ground.
The new manager had to absorb five of us into the team. It was difficult because they were already bonding for some time, the product was in general availability, and we were like guests asking for a space to sit and things to do. The irony of it all was that I was supposed to reframe the prior problems of supply chain security which we did previously but to also fit them into the main product. The communication pattern was also different and we tried to find our place in the new structure. Manager and I were trying to understand the scope of this supply chain problem and the gravity that the prior team termination meant for us and for the idea in general. Did we have to do the same thing but using different technology or should we just stop working on it? It was ambiguous with little customer led signal but just some ideas from individuals saying it needs to be materialised. That was another important realisation in big tech where some things would not make much sense and inertia alone can keep you afloat. The big question is if you just need to abandon it and jump ship or keep moulding it into something else.
A few years had passed, I was promoted to another internal level without a change in title. Yes, the thing got traction eventually. I was doing some analysis on my prior communications, documents that I created and ideas that I “drove” to justify the request to promote me again. When reflecting upon my prior work, I see the main differentiator from small companies is the need to keep throwing the ball until it hits the net, until some idea that you have or you adopt resonates with the people in the company. When things resonate you see more comments in the docs, other teams integrate with your service and join the calls. When that happens the software gets built and rolled out to datacenters globally. In small tech you just roll it out and talk to the customers or start seeing cash flowing and then can turn it off if necessary. Some might argue you should be able to do that in big tech as well but there are compliance checks, monthly business reviews and more points where the idea will be questioned, there needs to exist a wider support for it before you ship. When you seek support the idea might get diluted or it might be adopted by someone else who is not as excited about it. I think this is where politics starts and you need to know how to navigate it without making people angry and making sure you are still working on the things you want to do. All of this feeds into the promotion cycles and those in turn should motivate you to move forward and create impact to earn even more. You do not really care about the promotions in smaller companies, you might want a better title to show later on your CV but there is a limited number of them anyway. What you want is to sell more and increase the pool of available bonus or let some investors inject cash and convert your paper money. Actual sales figures and their relationship to your remuneration are much further away when in big tech if you are in engineering.
“AI AI AI” internal communication is beaming, our ship (the metaphor for the company) is changing course and everyone is waiting until we reach the golden coast. We have mountains of tokens to burn, access to the latest models, the internal tools are being changed. I’ve seen people burn xillions of tokens, what do they even do with them? I am adopting the tech as well and see the potential. I am sceptical at the same time but when you prepare the AI context well the thing works and gives you what you ask. Would I have access to these models and could I use that many tokens in a small tech company? I do not know. I am pretty sure we’d have Claude Code subscription or similar. Up until now I could compare big vs small but it is difficult to imagine how different the LLM exposure is and if it changes things. These tools did not exist when I was in a small team a few years ago. I’ll get back to it in some time to overview the effects and differences.
These are just some highlights that I remembered. I do not try to make a case if these should be the most important ones, just sharing some.