Thiago Julio

Software development

PT EN

Short, medium and long term: how I plan as a tech lead

leadership architecture roadmap

TL;DR - I split technical planning into three horizons: the next weeks (stability and support), the next months (migration and tooling), and the next year or two (architecture and team). Each horizon has its own goals and its own way to measure progress, and I review them every month.

The hardest part of being a tech lead in a startup isn’t any single technical decision. It’s that everything feels urgent at the same time: a customer bug, a migration halfway done, a new feature the sales team already promised. Without a plan, the loudest thing wins every day.

So I wrote down three horizons and what each one is for.

flowchart LR
  Short[Short term: weeks] --> Medium[Medium term: months]
  Medium --> Long[Long term: 1 to 2 years]
  Short -.- S1[Stability, support, small wins]
  Medium -.- M1[Legacy migration, tooling, review process]
  Long -.- L1[Architecture, scale, team]

Short term: weeks

  • Keep production stable. Fewer critical tickets month over month.
  • Support rotation working, so most of the team has focus time.
  • Quick wins the customers can feel: faster screens, clearer errors.

How I measure: number of critical tickets and how long they take to close.

Medium term: months

  • Keep moving flows out of the legacy, one at a time.
  • Finish the tooling upgrades and keep them current every month.
  • Code review standard and the review bot as the normal way of working, not a new thing.
  • Tech debt with its fixed share in every sprint.

How I measure: how much of the traffic goes through the new code, and how many debt issues close per month.

Long term: one to two years

  • The legacy backend switched off, or reduced to something small and stable.
  • Architecture ready for customers ten times bigger than today, mostly through queues and independent services.
  • A team where I’m not the bottleneck: more people who can review, deploy and decide.

How I measure: this one is harder. I look at whether decisions still need me to happen. If they do, I haven’t finished my job.

Reviewing the plan

Once a month I look at the three lists. Things move between them, some get dropped. The plan is not a promise. It’s a way to say no to the urgent thing that doesn’t matter, and to explain why.