What do I want to know about myself ?
I'm feeling out of sorts lately. Acting out more frequently than I use to. I realize most of the time, the reaction is not really to what had just happened. Its more about the unresolved emotions I still have from an encounter before.
Its important for me to understand these things because it bothers me, and when I act out, I often make decisions that I regret, that my future-self wouldn't agree. And these feelings are generally negative and damaging to my mental well-being, if I have go through them, then they have to be validated.
I always think that all of these things went back to the session when he told me about the bunch of candidates to be let go.
But why did it not affect me this much in the previous round of lay offs?
I think its because the targets are getting closer to my role proximity. And also because this time, he is not telling me after it happened, after everything is set and cast in stone. He is telling me, that internally, they have decided that these people will be asked to go in the next wave, in the coming year.
This knowledge, of knowing what is going to happen to people in the near future, is conflicting with so many areas in my work life that I have no of way of resolving or handling them.
One - I empathize and I wish that if this were to happen me, someone would signal me as early as possible so I have time to look for the next employment. But I can't extend this courtesy to them. Why? Because its the fucking hunger games. He said, if we spook any one of them, we might be put into the list as a replacement headcount to let go, if they leave on their own accord prematurely.
Two - I start to observe and feel insecure because the list of people to let go, is not cast in stone. Any kind of attention and work these people are given, I'd start to think that the list might have changed.
Why do I not trust him to do everything he can, to stop me from getting on the list? Because, deep down, I know didn't get to contribute to our components as much anymore in the last one year.
Year 1 was building data-comparer and getting myself familiar with the existing components we had back then. I was happy with myself because it was my first spark application. The demos were great and I get to see the effects of this in our day to day.
Year 2 was building AG2. It very beginning of GenAI era in the corporate environment. We started with building an API service in AWS to manage dynamic parameters for surveillance and this eventually extended to building a Knowledge Graph that would work with an existing LLM model off the shelve as RAG. It was the best year for me. Working with the interns to build POC on hosting the model, building various graph DB, working Langchain to hook up the LLM model with the graph.
It didn't take off eventually because, well, the regulations for accessing LLM models are not well established yet I believe, therefore it took them a while to democratize. Until then, there was a lot red tape around getting access to models just for POC in dev. There were a lot of politics as well in the LOB - lead architects come and go - the last one that stayed - decided to abort this project altogether.
Year 3 was building data-sentinel. It was great year too. Its my second Spark application, but on a much larger scale. We were set out to decommission the existing components because he was so upset with the on-going politics and decided we should just build from scratch. The product design was great because he was designing in hindsight. I took a lot of time to comb through the existing components, tried to pick up genuine feedback vs politics, understand the use of Spark - I went through different books on the use of Spark, understand the implementation of rule engine and plugins. It was fascinating. Requirements are generalized from top down because he really understands surveillance ops very well - the product ended up being very versatile and extensible.
Year 4 was re-enforcing data-sentinel. Onboarding and productionizing more use cases onto data-sentinel. There were a lot of politics and competition. Only after presenting to the global team and getting very hostile feedback and challenges - did I realized that I was building the same application that the strategy team, led by the lead architect, is building. The product design is definitely original but I guess instead of collaborating with him, the strategy team took his design and build an equivalent of data-sentinel on their own. And I think everything from here on, is about competing with Strategy team for him.
It was a difficult year. It was hard for me to go toe to toe with the strat team because each one of them has easily over 10 years of experience. I can read all I want but without experience - its hard for me to compete. For example, we don't have any use cases to run on clusters because the data size of the surveillance my team is responsible for does not require this setup. Things like that becomes my blind spot.
It was hard for him also because, I believe, his immediate line manager wasn't supportive of competing with strat. Instead, they expect him to pick the business aspect and compliance user management, again because of politics. We are hired as software engineers, but because of this distorted perspective over the years, that getting closer to the business is the ultimate "winning" strategy - I think they are not genuine nor do they have the capabilities anymore in developing people in the technology direction. Ironically, this how we lose so much control in the past one year and how the strat team came into power.
So we lay low for a bit, focusing on business, engineering was stale. I tried to progress in the business direction but I'm a builder - this was really awkward for me. Eventually, GenAI enablement caught up. We can access LLMs without red tapes or having to worry hosting infra, gpu and memory and things like that. All of these things are now abstracted away from us, handled by a central team. Models are easily accessible with unlimited capacity.
This in turn helped him to advance very much in the business aspect. You see, the challenge in business, during the pre-GenAI era, was interpreting their requirements and their endless and relentless questions on the scope of surveillance, and then matching with what we have engineered - to either give them assurance or declare that something is indeed missing and recommend what to fix. With superior reasoning powers and unlimited capacity, if we've done proper context engineering, majority of the translation between business and engineering can be handled by the agent.
This is definitely does not cover everything that had happened. I haven't really talked about the demo of semantics+lineage POC that didn't went well. I didn't talk about what my friends told me that their manager didn't think he was being fair to me because he didn't really fulfil his managerial role in giving me exposure to the right people. And I didn't talk about how I intuitively identified moments that would concur to this. I also didn't talk about how he is suggesting to get one of the candidates to be let go, to do the same work he just assigned me because he wants win the competition.
But the recounting so far is enough - I think my insecurities are mainly fueled by following:
the management's distorted perspective on software engineers - and its going to be infinitely enlarged because of GenAI. The chances of me getting technology related experience to become more advance now is greatly reduced
the on-going retrenchment because of GenAI
we have no engineering work this year and my contribution to use of GenAI for the business is very minimal
I don't trust that he will think that keeping me is more important winning the competition. Now that he is free of burden from managing business full time - he wants to compete again.
The way I know to be good at my job - analyzing, studying, experimenting by building - these are not true anymore.
Are all these observations valid? What do I do next?
Comments