Dr. Cecilia Dones, at 3 Standard Deviations nominated for the #WomenTech Global Awards 2025! Share to raise awareness! #womenintech #womeninstem #womenwhocode #WTGA2025 @womentechnet url:http://www.womentech.net/nominee/all/all/147964
Most PMs chase precision, but it often hurts the experience.
I use a food-tracking that uses AI to measure
It’s not exact, but neither am I
What matters isn’t perfect data, it’s progress
Meet people where they are and give them one step that moves them forward
Every company today has a junk drawer of infrastructure, tools
and a very specific point of view
It’s nearly impossible for a product solve it all
I'm guessing the forward deployed engineer is a way to cover the gap
What am I missing?
I keep seeing forward deployed engineer roles, guess it's a thing
Every new role emerges to solve a need
Sometimes real, Sometimes perceived
Palantir seems to have kicked things off here
embedding engineers directly with clients to deal with mess
Your team’s growing, slower
Projects drag. Decisions stall
That’s Scaling Sludge: the friction success creates
You don’t lose to smarter rivals, you lose to ones who can move.
The cure? Clarity
If your training data misses edge cases (or worse, is mislabeled) you’re going to have problems.
Good training data is:
Accurate
Relevant
Consistent
Comprehensive
Balanced
Plentiful
Clearly labeled
No shortcuts. No snakebites either. We warned others and rerouted.
What finding a nest of copperheads taught me about training data
True story but kind of silly.
We’re hiking to a climbing spot. My brother spots movement in the brush.
He’d seen enough examples to recognize the snakes.
AI runs on examples too.
Your model is only as smart as the data you give it.
AI is reshaping how we build, code, test, and deliver.
Tomorrow I’m sharing lessons from the frontlines:
what’s changing, what’s at risk, and how builders can evolve.
Register at https://f.mtr.cool/cnunczwqrj
Ever been in a 45m meeting with no decision?
The direction disconnect
when strategy and daily choices don’t align.
Simple rules that guide choices, cut swirl, and speed progress.
“Make them tell you no.”
My mom said this every time I doubted myself.
A job. A team. A talk.
Don’t count yourself out before you try.
HBR found men apply at 60% qualified, women wait for 100%.
Apply. Ask. Pitch.
You don’t have to be ready. Just go.
I’ve been helping people sharpen their one-liners.
What’s working?
1) Let it rip. Half-formed ideas and vibes welcome.
2) Get someone outside your bubble.
Supportive play leads to real clarity. And yeah, it’s pretty fun.
Here’s what works:
• What the problem looks like
• How we solve it
• What the customer gains
• How we plug into what’s there
Show how it hits the number → you earn attention.
PMs, designers, eng: go on a ride-along. Feel the stakes.
#buildingtomarket
Confession: I bombed a pitch to my own sales team.
I talked design wins, engineering magic, happy clients...
Got nothing.
Sales lives and dies by the number.
If you don’t help them hit it, you’re noise.
I had to learn to tell a different story.
Location, location, location isn’t just about real estate, it matters in the cloud too.
Where your data lives impacts compliance, cost, and performance.
Slow down your users
Rack up transfer fees
Break laws you didn’t know applied
A “region” is the location of your data center (think US East)
Teams are reporting barriers to collaboration
49% Competing incentives and internal politics
43% Lack of leadership support or buy-in
35% Conflicting priorities across teams
Data from Atlassian State of Product 2026
Leaders should be digging into this and taking those barriers down
We've got leadership problems on product teams
80% of product teams don’t include engineers in ideation, problem definition, or roadmapping.
Coming in too late to contribute meaningfully, especially on AI features that need early feasibility input.
The “heft test” is 💩.
Impressive decks from top firms that do nothing but make people feel better while burning budget.
Real change comes from clear direction, real engagement, and visible progress.
To make big things happen, you need clarity, engagement, and the discipline to test and adapt.
AI leaders say PMs need to write evals
Evals are how you define “good”
Define success in clear, testable terms
Align your team around outcomes
Avoid "it kinda works?" moment post-launch
https://f.mtr.cool/bolppzledc
#AIProductManagement
You can’t build in a vacuum and expect relevance
You can’t build AI-powered products without fast learning
And get this:
49% don’t have time for strategy.
49% can’t fit in data analysis or metrics.
We need to have a serious talk about time management gang
Source: Atlassian State of Product 2026
You’re not ready for AI if you're not even iterating
84% of product teams are worried their product won’t succeed
Only 31% are doing rapid iteration.
This post breaks down:
How to spot the root cause of performance issues
What each factor looks like
What to do before blaming apathy
Read full post: https://f.mtr.cool/byiiejsysz
#leadership #teameffectiveness #productmanagement #workculture #growthmindset
Struggling with team performance?
Harvard’s Capability–Motivation–License (CML) framework points to three common reasons teams stall:
Lack of skills
Low motivation
No permission to act
But there's a fourth factor that quietly shapes them all: Direction.
Builders are evolving. Are you?
Next month, I’m speaking at an event hosted by Agile New England on how AI is reshaping product, design, engineering and delivery.
I'll share lessons from people building AI products and building with AI from OpsCanvas and beyond.
More information to come
Ever spin up a server, give it extra memory and CPU “just to be safe,” and never turn it down?
That’s overprovisioning and it’s draining your budget.
In this episode of What the Cloud:
What overprovisioned resources are
Why they waste money without adding value
Got a cloud question? Ask away
So why in the cloud are we still asking engineers to tag resources, pretending tags are accurate, and building governance on broken data?
Feels like we should be past this. Tools can organize, categorize, and contextualize automatically.
#noobquestions
Why is tagging even a thing? Didn’t we already solve this in the rest of tech?
I’m new to #cloudops and #finops and I’m confused.
You don’t tag your photos. Google Photos figures it out.
You don’t tag every Starbucks charge. Rocket Money or FreshBooks categorize for you.
MIT says 95% of generative AI pilots are failing
The technology usually works. Integration doesn’t.
Tools don’t adapt to company processes
Companies don’t adapt their processes to the tools
Adoption requires a willingness to rewire how work gets done
The 5% who figured this out, did it together
Been talking to engineers about #IaC
Powerful, often too complex
Benefits: consistency, version control, speed, recovery, collaboration
Challenges: complexity, learning curve, secrets mgmt, costly testing, upkeep
I'm thinking AI could help
Always looking to interview people working in this space
What do I do when I don’t know the #strategy?
Many orgs are bad at strategy.
If you’re lucky, it exists but isn’t communicated well.
If you’re unlucky, it’s bad
Look for clues
Find a senior person who knows
Take a guess and get a gut check
Default to basics: make money, save money, reduce risk
