Professional year in review
I got a lot done professionally in 2025.
For most of the year, I was a backend engineer on the recommendations team at Hinge. We started the year by rolling out my big 2024 project, which was adding Elasticsearch to power new candidate generators in Hinge’s recommender. The initial launch was successful: its p99 was 80% lower than our previous Postgres-powered version, and was much more maintainable. For the remainder of the year we added more and more features using this stack, and now it’s load bearing for a bunch of product wins. Along the way we learned a lot about scaling Elasticsearch clusters, so I’ll try to write a blog post or a conference talk explaining our approach!
By the end of the year, I became the backend tech lead of a recommendations-adjacent team called “Matching.” So now I’m spending my time doing the usual TL dance: refining new ideas with product managers, guiding our tech implementations, and then project leading and IC work when everything is humming smoothly. I’m really excited by our roadmap, so I’m hoping 2026 will be a good year as well.
Multiple people have told me that I bring “fun dad energy” to work every day. I’ll take it.
Fun dad energy
My daughter is 2.5. I love hanging out with her. She tells jokes that catch me off guard and make me laugh. She loves stories and playing with words. She’s trying to guess the first letter of words based on how they sound. She even gets it right sometimes. She’s starting to develop real friendships with the kids at school. For the first time, she sometimes wants to just play by herself for 10 or 15 minutes. 2024 was all about her becoming an individual we could talk to. 2025 was all about her growing in complexity. I’m so excited to see what else changes this year.
We tried to develop her athletic side by enrolling her in soccer. Every Sunday, we’d bring her to a field near our apartment. She’d ignore all instructions from the coach. We needed to intensely micromanage her to even kick the ball. Halfway through every class she would jailbreak by sprinting across the field away from the lesson. Her coach was great with kids, which is probably the only reason my daughter was excited to go every week. But a switch flipped after a few months of attendance. She started listening to her coach. She gleefully did everything the coach wanted. And she actually stayed in class instead of trying to flee. We’re thinking of enrolling her again for the spring. I think she’ll actually want to go.
Much of the rest of the year was pretty mundane. Her sleep schedule stabilized. She’s sleeping more than she used to, and so are we. Potty training was hard but achievable. Giving her a balanced diet was hard but achievable. That’s a good summary for the year in parenting: hard but achievable.
New apartment
In 2024, our building installed heat pumps. It was a massive project that lasted most of the year, finally ending in November. The 1 bedroom was getting too small for us, so we moved into a rental in the middle of the year. We wanted to sell our apartment but had to wait until the construction finished. And then the unexpected happened: my next-door neighbor put her apartment on the market. Against all odds, she accepted our bid. We moved in at the end of 2025 and I’m living in two apartments next to each other. I still can’t believe it. For the first time in my adult life, I have enough space.
Our plan is to combine the two units into one next year. This will be a nightmare. However, it’s 2027’s nightmare.
We have one advantage: the units might have been combined in the past. Most of the wall between the apartments is a thick load-bearing masonry wall. But I knocked on the walls between the units. There are 2 doorway-sized areas that sound hollow, and are covered in sheetrock instead of plaster. I dug in with a long screwdriver and it sounds like there’s a different brick wall between the two (instead of the 12-ish inch thick load-bearing wall). I’m hoping that these are doorways that were bricked over, and that we can just take the brick out and live happily ever after. But I won’t know for sure until we take the sheetrock down. Again, 2027’s problem!
Newsletter
My daughter had a more predictable sleep schedule this year. So I was better rested and had more free time. I decided to write a newsletter at clientserver.dev. My initial writing prompt was “Money Stuff for tech.” I committed to writing two posts per week. So I focused on current events in tech. Since my wife went to bed at 10:30 and I went to bed between 12:00 and 12:30, I had 3-4 hours to write each post.
I wrote two posts per week in the beginning. Over 6 months I got to 268 subscribers. Holy cow, people wanted to read it! This made writing easy: every single post had a guaranteed audience. However, the newsletter took up ALL of my free time. I worked around the clock. I pushed my bedtime later and later to get issues out. I was exhausted. I became more stressed. I become emotional when small changes in my schedule took away time from the newsletter. My wife started begging me to take time off. Once she started trying to intervene, I was like “this is a really bad sign.” I stopped working on it in July. Next year, I’m going to start publishing it again without a regular schedule. I think writing is too valuable and I don’t want to throw the audience away. bitlog.com will continue to be my own personal writing, and clientserver.dev will continue to be my hot takes on the news.
Having a short stint as a tech writer made me realize a few things:
- Sourcing stories on a regular basis is extremely hard.
- The good publications are incredible. Shoutout to outlets like The Register that produce high-quality tech journalism. They do an incredible amount of research and writing on a timeframe I can’t fathom.
- Most publications are absolutely horrific.
For example, journalists apparently don’t have time to read anything. They just get a prompt and write and publish. I’d occasionally find things that were (a) mass reported, and (b) trivially provable to be wrong. For example, many outlets reported that Salesforce would not hire software engineers in 2025. But anyone could go to Salesforce’s job page and see dozens or hundreds of software engineer job postings. And then you’d read Marc Benioffs’ statements and interviews on the subject and you’d realize, “oh, they actually said they are keeping their engineering headcount stable. All of these people are reporting the wrong thing.”
Gaming
As I said in the newsletter section, I’ve had more free time this year. After my newsletter stint, I’ve picked up more games in my free time.
My #1 game of the year was Hollow Knight: Silksong. What else is there to say? Great soundtrack. Great bosses. A fun moveset on Hornet. If I had any criticisms (besides Bilewater), it’d just be the absence of a challenge boss like Radiance. It feels like a missed opportunity given Hornet’s versatility and mobility.
I also sunk almost 100 hours into Blue Prince. Which is a bit weird given that I did not like it! The game is like 40 hours of gameplay spread into 300 because of RNG, and my life is too busy to allow some dude to waste entire days of my life like that. On the plus side, some of the puzzles were genuinely enjoyable. Color was also starting to be a factor in puzzles. I am colorblind with protanopia, which is a severe color deficiency. This game was so subtle and tricky that I never knew whether I was missing some obvious clue, like “this realm’s color is only exposed in one place, where you have to notice that a postcard is tinted a specific color.” I also understand that color continues to be more and more important as the game progresses.
Spoiler paragraph for anyone curious about how far I got: I was playing with 0 hints or outside help. I love word puzzles and had a lot of fun decoding the Baron’s Bafflers, and needed a few visits to the gallery (and a hint from the classroom) to solve all of the painting puzzles. I was clearing out the tunnel, I had gotten most of the way through the 8 doors of the realm puzzle (probably close enough to brute force the remainder). I uncovered the CASTLE puzzle and (correctly) had a few clues that I thought were part of the puzzle. I also unlocked the throne room and figured out how to steal the crown, but wasn’t sure if they were connected in any way. I was collecting trophies so that I could unlock the blue tent to see if it did anything or just wasted money. I finished the classroom quiz and looked at the giant pile of hints I hadn’t processed yet in my Notion doc, and the list of things that I was waiting for perfect RNG to do, and just couldn’t bring myself to try the next thing. So I called it quits and looked up the remainder of the story and major spoilers (Thanks, FuryForged!). I’m not sure if I could have 100%’d the game without help, but I didn’t have the time necessary to find out. I also didn’t like the story or the lore, although I did appreciate that once you understand the early through late game, it’s a story of how a spoiled brat truly earns the right to call the house his own.
I’ve also rediscovered the joy of just messing around in party games with my friends! Recently we’ve been playing “Golf With Your Friends” and “RV There Yet” and having a good time.
Looking ahead
My goals for 2026 are pretty simple:
- Get in shape
- Take more time off
- Find more time to write
See you next year!

