Using AI to Get Healthier: An SRE's Approach to Personal Reliability

Most of my waking hours are spent thinking about reliability. How do we keep systems running for apps that bring in millions of dollars and serve millions of users? How do we automate away the toil, detect incidents faster, and build infrastructure that doesn't fall over at 2 AM? I've been using AI to help build that infrastructure, generating Terraform, analyzing logs, accelerating incident response. It's become a core part of my workflow.

But at some point, I had a thought that honestly should have hit me sooner: I'm spending all this energy making sure systems stay healthy, but I'm not putting that same energy into making sure I stay healthy.

The Wake-Up Call

When you work in SRE, you get comfortable with data. You monitor everything. Dashboards, alerts, SLIs. If a service is degrading, you know about it. But I realized I had no real visibility into my own health metrics. I had a Samsung watch collecting data that I barely looked at. I had a treadmill collecting dust in the corner. I was optimizing uptime for production systems while my own "uptime" was questionable at best.

I know better than anyone that you can't improve what you don't measure. So I decided to treat this like any other reliability problem.

Enter Google Gemini

I have the Google AI plan through Verizon, which comes out to about $10 a month with the discount. I was already using Gemini for work tasks, so the barrier to entry was basically zero. I decided to see what it could do with personal health data.

I went all in. I gave Gemini everything I could about myself:

• My daily routine: wake time, work hours, meals, when I typically crash on the couch
• My diet: what I eat on a regular basis, portion sizes, snacking habits, the good and the bad
• Samsung watch data: steps, heart rate trends, sleep patterns, active minutes (or lack thereof)
• My goals: where I want to be, what I'm willing to commit to, and what's realistic given my schedule

The more data I fed it, the better the output got. It's the same principle we follow in SRE: garbage in, garbage out. If you want meaningful insights, you need to give the system meaningful data.

The Plan Gemini Built

Based on everything I provided, Gemini generated a weight management and fitness plan tailored specifically to me. Not some generic "eat less, move more" advice, but an actual structured plan that accounted for my work schedule, my eating patterns, and my current fitness level.

One of the most motivating things Gemini did was generate a spreadsheet with weight goals and timeframes. Seeing the targets laid out week by week with projected milestones made the whole thing feel real and trackable. It turned a vague idea of "lose weight" into a concrete, measurable plan. Having that spreadsheet to check against keeps me accountable.

A few other things stood out:

• It identified patterns I missed, like how my worst eating habits correlated with high-stress work days
• It was realistic. It didn't tell me to wake up at 5 AM and run 5 miles when it knew I work late
• It gave me incremental steps, small, achievable changes instead of a complete lifestyle overhaul
• It incorporated what I already had, specifically the treadmill that had been sitting in my home barely getting used

Dusting Off the Treadmill

I've had a treadmill for a while now. If I'm being honest, it was more of a clothes hanger than a piece of exercise equipment. But Gemini's plan incorporated it as a daily habit. Not hour-long marathon sessions, but consistent, manageable walks and incline work that fit into my day.

The key was making it non-negotiable. Every day. Not "when I feel like it," not "if work isn't too crazy." Every day. Even if it's just 20-30 minutes. Consistency over intensity. Gemini helped me see that showing up daily at a moderate effort beats crushing yourself three times a week and then burning out.

It's a lot like how we think about system reliability. Steady, sustainable performance beats heroic efforts during outages.

Getting Outside and Working Out at Home

Now that the weather is finally getting nice here in Minnesota, I've started going on outdoor walks too. I try to get outside at least three times a day. It breaks up the workday, gets me away from the screen, and honestly just feels good after months of being stuck indoors. Between the treadmill and the outdoor walks, I'm moving way more than I was even a few weeks ago.

I've also incorporated at-home workouts into the routine. I can't find the time to go to a gym, and honestly a gym membership is another expense I don't need right now. One of my goals through all of this is to save money. Between the Gemini plan, the treadmill I already own, bodyweight workouts at home, and free outdoor walks, I'm keeping costs basically at zero beyond the $10 a month for the AI plan. No gym fees, no personal trainer, no expensive meal kits. Just data, consistency, and effort.

Treating My Health Like a Production System

The parallels between SRE and personal health are honestly hard to ignore:

• Observability: I started actually looking at my Samsung watch data. Steps, sleep quality, heart rate zones. You can't improve what you don't observe.
• SLOs for myself: I set personal targets. Minimum daily steps, minimum treadmill time, a calorie range to stay within. Not perfection, just a target to trend toward.
• Incident response: Bad days happen. I eat garbage, I skip the treadmill. The point isn't to never fail. It's to recover quickly and not let one bad day spiral into a bad week.
• Automation: Gemini helps me meal plan and adjust my routine without spending hours researching nutrition. The AI handles the toil so I can focus on execution.
• Iteration: I check in with Gemini regularly, update it on progress, and let it adjust the plan. Just like we iterate on runbooks and dashboards, the health plan evolves with new data.

Why AI Made the Difference

I've tried getting healthier before. Everyone has. What usually killed it was decision fatigue. After a full day of making complex engineering decisions, the last thing I wanted to do was research meal plans or figure out an optimal workout split. I'd default to whatever was easiest, which was usually doing nothing.

Gemini removed that friction. I dump my data, tell it what's working and what's not, and it gives me an updated plan. It's like having a personal trainer and nutritionist available 24/7 for $10 a month. No scheduling, no judgment, just data-driven recommendations. And it fits perfectly with my goal of keeping costs down.

Where I'm At Now

I'm still early in this journey, but the treadmill is getting used every single day now. That alone is a win. My Samsung watch data is trending in the right direction. I'm more aware of what I eat and when. Most importantly, I'm actually consistent, which has always been the hardest part for me.

I'm not writing this because I have some incredible transformation story to share yet. I'm writing it because the decision itself mattered. The decision to take the same tools and thinking I use professionally and apply them to my own well-being.

The Takeaway

If you're an engineer who spends all day making sure other people's systems are reliable, take a minute and ask yourself: am I treating my own health with the same level of care?

You already understand observability, data-driven decisions, and iterative improvement. You probably already have the tools: a smartwatch, a fitness app, access to AI. The hard part isn't knowledge. It's deciding that you're worth the same investment you give to production systems every day.

I decided I was. The treadmill is no longer a clothes hanger. The Minnesota weather is finally cooperating. And Gemini is no longer just for writing Terraform.

I'll be posting updates on how this goes. The spreadsheet has the targets, the plan is in place, and the work is happening daily. Stay tuned.