I've been in tech for over 10 years, across five companies: fintech, SaaS, and my own startups. Every time I update my resume, promo case, or LinkedIn, I hit the same wall: I genuinely can't remember what I did.
Not because it wasn't meaningful. Because nobody wrote it down as it happened. Ask me what I shipped in Q2 last year and I've got nothing, until I sit with it and slowly reconstruct. That's career amnesia, not a personal failing.
And even when I do remember, the resume format kills the part that mattered. "Managed cross-functional team to deploy new architecture" tells you nothing about the problem, the constraints, what broke, or what I learned fixing it. A resume doesn't just compress your career. It strips the context, and the context was the point.
Now add AI: a polished, keyword-optimized resume takes 30 seconds, whether you did the work or not. The format that already threw away the substance is now trivially fakeable too.
So what's left? Not what you say you did, but a dated, ongoing record of what you actually did, with the context intact.
What's already out there
A few tools circle this problem. Each solves a slice:
- Dayflow: open-source, local on your Mac, turns screen activity into a daily timeline and standup summary. Great for "what did I do today?" Days and weeks, not a career.
- bragdoc.ai and similar: pull GitHub commits and PRs into an AI brag doc for review season. Developer-only, and built for one moment a year.
- ActivityWatch: long-running open-source activity tracker. Strong on raw data, no narrative.
- Trajecta: closest in spirit, a permanent record of professional impact. Still manual entry. You're the one writing it.
Each nails a piece: local-first privacy, developer activity, raw tracking, long-term framing. None puts it together: passive where possible, guided where it can't be, non-technical roles included, built for years instead of days.
What I'm building
That's the gap I'm building into with EchoWork.
Where it can, it pulls from where the work already lives: commits, tickets, calendar. Where it can't (enterprise, IP-locked environments), it doesn't scrape or spy. It notices you had a three-hour block on "system migration" Friday and asks what you decided and shipped. Either way, it becomes a real monthly rollup: not a bullet point, but the actual story, dated and preserved.
It runs local-first, with your choice of LLM, so none of it has to leave your machine.
It's not built to replace LinkedIn or hand your data to recruiters. It's for you first, so when review season, a promo case, or a job search comes around, you're not starting from zero. Sharing a slice with an application is optional, not the point.
I'm building the first version for myself right now, testing whether I'd actually use it daily before deciding what it becomes. I think the resume's days are numbered. What replaces it is proof, with the context still attached.
Top comments (1)
What would you want captured first: calendar blocks, tickets, or the "what did we decide" moments?