amuino.bsky.social
Can’t fix everything I’d like to.
Currently VP of Engineering at Cabify.tech
Galician @ Madrid.
82 posts
89 followers
61 following
Regular Contributor
Active Commenter
comment in response to
post
I'm not really sure how to reconcile these things – another big example that I haven't been able to square in 5+ years remote is how to properly train junior staff without the ability to shoulder-check with a gaggle of intermediate and senior staff.
comment in response to
post
It also seems that remote and distributed are often confused.
When a company hires 100s of kilometers away from any office, there is no RTO possible, no hybrid setup that works, etc…
People can only Return to Office if they ever had one.
comment in response to
post
This definitely happens.
We’ve lost a few persons that loved the job and the company, but needed the in-person interactions and moved to hybrid/on-site companies.
We were their first fully remote engineering team and did not have nearby offices… so RTO for them meant changing companies.
comment in response to
post
The most surprising part, the efficiency. We achieve all of that with ~15 persons crafting products around k8s, that are then used by over 150 other engineers.
I’m super proud of the team and what they have achieved. This is how high performance feels, but we use DX to have more than feelings.
comment in response to
post
Deploying services, exposing them to the internet, securing acces via our own IdM, bridging clusters, autoscaling, rate limitting, CVE scan choke points, out of the box monitoring… that’s all solved by Platform products built around k8s (some old enough to have wrapped Nomad back in the day 👴🏻).
comment in response to
post
In the mean time, Cabify Platform Engineers have built a collection of powerful abstractions and guidelines to make the most of out k8s. Few Product Engineers need to get into the nitty gritty of k8s.
comment in response to
post
Para Dev&Tech me salen solo cosas de 2023 y nada en “Upcoming” 🤔
comment in response to
post
Gracias!
SCBCN es la única que tenía en mente! ✅
comment in response to
post
Hornillo no, pero un walkman más viejo que ese sí que hemos sacado.
comment in response to
post
But there’s a catch. Lisanne Bainbridge highlighted it back in 1983 in her paper "Ironies of Automation": the more we automate, the more we deskill.
comment in response to
post
Por cierto, mientras esto se arregla: gpt-o4 es MUCHÍSIMO más rápido que Claude 3.7 sonnet.
comment in response to
post
Balancear el corto y largo plazo de una organización no es trivial.
“Best Simple System for Now”, aplicado a organizaciones
dannorth.net/best-simple-...
comment in response to
post
Las empresas contratan cuando les duele, con lo que muchas buscan la “I” que les quita el dolor de hoy. En general, los perfiles “M” son más caros y raros, los budgets son apretados y las justificaciones difíciles (coste de oportunidad de esperar a una M “cara” vs una I barata y ya disponible).
comment in response to
post
All in all: Cursor lacks autonomy (thankfully?) and ability to abstract instructions to create a richer domain… but otherwise is like managing work for an IC. Someone very technical without understanding of the actual problem domain but without any ego, so accepts any guidance.
comment in response to
post
From time to time, I suspect or notice complexity has accumulated. For instance, when the code style checker complains about too many arguments to functions. Here I step in and suggest a refactoring (cursor seems bad at detecting abstractions).
comment in response to
post
When I want a new feature, I ask for a refactoring and I even get a data migration script when the database schema needs to change.
comment in response to
post
I asked for some quality checks, and I see Cursor running tests, checking code coverage does not go down, following code style…
I check the user experience does not degrade (my experiment inckudes an interactive web tool with drag and drop and not trivial stuff)
comment in response to
post
I’m getting started with Cursor and I’m on my honeymoon phase. At this stage, I’m barely looking at any code and I’ve been happily “vibe coding” stuff looking at the outside effects.
comment in response to
post
Serendipio. Os leo aquí hablando de M-shaped y se me cruza esto un poco después…
spin.atomicobject.com/broken-comb-...
comment in response to
post
PS - we have a great relationship & regular one on ones. But I've found some engineers don't think the stuff they do is worth mentioning. Even though it is! They view the work as "oh, well it needed to be done" or "they needed help so I helped". Not realizing how valuable and unique that is.
comment in response to
post
As a former shit umbrella turned ops person, if you need a shit umbrella, something is terribly wrong in the org.
I will hardly fault a manager for being one because that's what they know they can do right now in this specific context.
comment in response to
post
Couldn't agree more
I just gave a talk where I showed my version of org health (encourage others to do the same), where teams who don't fight at all are at the lowest end of org health- false harmony just leaves issues to simmer
comment in response to
post
💯 Lo de siempre: entre otras cosas, falta "outcome over output", "eficiencia de flujo sobre eficiencia de recursos", el engaño de la "alta ocupación", foco en "mover la aguja", foco a secas, falta de feedback real de usuarios/clientes/negocio, falta de autocrítica e introspección seria, etc. 🙃
comment in response to
post
Y antes de lanzarte, puedes investigar más de CabifyTech, incluidos los salarios, en cabify.tech/handbook/ para ver si hay feeling.
comment in response to
post
Tenemos vacantes de SRE en Cabify para ese nivel de experiencia (lo llamamos L3). Trabajamos con AWS y GCP, pero no “entrevistamos de AWS” (ni GCP). Nos interesan más los fundamentos.
Te dejo link: cabify.careers/es/job/44015...
comment in response to
post
Se habla mucho de bitcoin y muy poco del cacao
comment in response to
post
Link obligatorio para los que estéis en desacuerdo.
cabify.careers/es/jobs?sear...
comment in response to
post
¿Puede ser que el enlace a la galería del 2015 apunte al 2014?
(He ido buscando mi dosis de nostalgia)