A Manager README is a good way to introduce yourself as an Engineering Manager. I first wrote one in 2021 — the act of writing it down taught me more about myself than the document ever taught anyone else.

I wrote the original version in 2021. This is the 2026 rewrite, and a lot has changed — not so much what I believe, but the job itself. The work I do now looks less like managing a team and more like orchestrating a system of people, tools, and increasingly AI. So this README has two layers: a thin one that changes often (how the role works right now) and a thick one that barely moves (what I value). I’ll keep updating the first. The second is the part worth sharing.

The original 2021 version is here. This replaces it.

Disclaimer: Nothing here should be read to apply to any other group or manager at Grafana Labs, my current employer. This is about me.

Why a README?

This is a user guide to me — my values, my operating principles, the things that delight and disappoint me. It is not a replacement for the relationship we build by actually working together. Think of it as a head start.

Discrepancies between what is written here and how I actually behave are dangerous red flags. Call them out.

What I’m actually accountable for

The textbook answer — set the team up for success, grow people, deliver value — is still true, so I’ll keep it short:

• Setting and holding the long-term direction of the team, and connecting it to what the company needs and what each person needs.
• Building a culture of learning, trust, and collaboration.
• Attracting, growing, and retaining people who do excellent work — and growing our leadership capacity.
• Making sure the team ships the highest-value work we can.

But here’s what the textbook answer misses in 2026. I increasingly see myself as an orchestrator. AI has changed the shape of the work. Everyone is building, faster, with a co-intelligence at their side. The bottleneck has moved from can we build this to what should we build, and in what order — so that’s where my attention has moved too. I spend far more time thinking about priorities and direction than about how any individual gets their work done, because there is now almost no limit to what can be done.

A second shift: I find myself teaching management skills to people who don’t manage anyone. When everyone is directing AI agents and coordinating work across boundaries, everyone is doing a version of management — delegation, prioritization, holding a thread of accountability. So a growing part of my job is helping ICs build those muscles.

Why I like this job

Honestly? I love getting things done without doing them myself. The orchestrator part — seeing results land, watching other people get things done, knowing I was the mastermind quietly behind it — that is deeply satisfying to me. That’s the real answer, stripped of the noble-sounding ones.

My north star

I used to write that I wanted to take a long break and have it make no difference to the team. That was aspirational. The reality is that it always makes a difference — and pretending otherwise undersells the strategic half of the job, which is already chronically undersold.

So the honest version: I want to make sure that when I’m not there, no one is stranded. Everyone has clarity on what to do, clarity on what matters over the long run, and clarity on who to reach out to when they’re stuck. Not invisibility. Just no single points of failure that run through me.

Observable management

Much of a manager’s work is what I call invisible-important — high-leverage and completely unseen, which breeds mistrust and misalignment. I try hard to counter that by making my work observable: sharing not just what I do but why, and defining the impact of my work in terms the team can actually see. I keep a living “What does Vardan do?” document with my rough time allocation, and I track team health openly. If you ever wonder what I’m spending my time on, the answer should never be a mystery. Ask, and if it’s not already written down somewhere, that’s a bug on my side.

What I expect from the team

In my ideal team:

• We trust each other.
• We hold each other accountable.
• We assume positive intent.
• We learn together and are never afraid to ask questions.
• We celebrate success together and are pragmatic about failure.
• We help each other reach excellence — which sometimes means honest criticism.

On conflict, my view has matured. I used to write “we don’t avoid conflict.” That assumed people will engage with conflict if you encourage them to. They mostly won’t — avoiding conflict is the human default. So my job isn’t to cheerlead for conflict; it’s to build the structure that makes conflict safe and routine enough that people actually engage with it instead of routing around it. Assume the conflict will happen. Assume people will dodge it. Then build the environment that helps them face it and learn from it.

My leadership philosophy

I don’t believe in one superior style. By default I gravitate toward:

Orchestrating

I think in systems and flows more than I used to. I’m less hands-on about how people do their work — they have AI as a co-intelligence now, and they don’t need me hovering. I spend that reclaimed time on direction, sequencing, and unblocking.

Empowering and guiding

I feel most accomplished when I help someone become their best self. I’ll share stories, learn out loud, and be present when people need it. Where I have a view on a better path, I’ll offer it — increasingly as concrete, tactical advice rather than pure open-ended coaching. In an AI-saturated world, people often need a clear next step more than another reflective question.

Individualization

I’m fascinated by what makes each person distinct, and impatient with generalizations and types. I’ll always work to find the balance between what a person needs and what the company needs.

Participating

I like being in the day-to-day — brainstorming, design discussions, the actual texture of the work. Not to control it, but because participating is how I learn and stay useful.

Values

These barely change. How I apply them changes constantly.

Fairness first

I optimize for fairness, in every game we play. This is not “I favor people over the business, faults and all.” It’s the opposite of soft: I insist the game itself be fair. Most systems are tilted by default — credit flows up, blame flows down, the transparent get more scrutiny than the opaque, the last-minute hero gets rewarded over the person who quietly prevented the fire. I work against that tilt. When I do, people tend to win — not because I put a thumb on the scale for them, but because a fair game was being denied to them in the first place. Invest in people fairly and the business benefits. That’s the whole bet.

Agency

This is the one I light up about. Autonomy is permission to decide how you do your work — I give it freely. Agency is bigger: it’s people standing for their problems, refusing to let something stay broken, going after it themselves before anyone asks. I don’t condemn its absence so much as deeply appreciate its presence. If you have agency, working with me will feel like a tailwind.

Trust and psychological safety

The foundational block. By default, I trust you. Safety, for me, is no longer just “I’ll protect you” — it’s structural: building an environment where hard things are survivable, where it’s safe to be wrong out loud, to disagree, to surface bad news early. Edmondson’s work on psychological safety sits underneath a lot of how I operate.

Transparency

I expect work and status to be visible — not for surveillance, but because I can’t do capacity planning, sequencing, or honest advocacy for the team if I don’t know what’s actually happening. And I hold the same standard for myself: that’s what observable management is about. One asymmetry I watch carefully — the people who are most transparent should never get punished for it with extra scrutiny. Openness has to be rewarded, or it dies.

Equanimity

A more recent and, frankly, more Stoic addition — it runs through my EM Meditations. I try to spend my energy on what I can actually control and make peace with what I can’t. I can translate a decision I disagree with without litigating it forever. I try not to let events disturb my judgment of them. This is aspiration as much as practice, but it’s genuinely how I want to lead.

What disappoints me

• Lack of accountability. I don’t tolerate finger-pointing or blaming others. If something went sideways, I want self-awareness about your own part in it — not a search for who else to name.
• Lack of transparency. Tell me the truth — about a project’s status, about me, about a hard problem. I can absorb a hard truth far more easily than a comfortable evasion.

I deliberately dropped “idling” from the old version. I’ve come to believe engineering joy is shifting from tight, craft-based feedback loops toward slower, outcome-based satisfaction as AI absorbs more of the direct problem-solving. Measuring busyness was always the wrong instinct; it’s even more wrong now.

Feedback

I give feedback as close to real-time as I can, and I’d love the same from you. I prefer to receive it in private; candid-and-compassionate is the best gift you can give me. Nothing you say in feedback will ever affect our relationship or your standing — full stop.

On candor: I’m honest, but I’m not “radically candid” in the say-everything sense. I might say three of the five things on my mind and let two go. That’s not me hiding — it’s judgment about which truths are worth their cost, and when. I’d rather land the things that matter than spend the room’s trust on the things that don’t.

Performance

Recreation is essential for creation — I don’t expect anyone to perform exceptionally all the time. I do expect exceptional work regularly. Performance, to me, is the combination of what you deliver, the impact it has, and how you work with the people around you. When that’s persistently off from what we agreed, I treat it as a red signal — and my default is to work with you to recover from it, not to give up on it.

On AI and the future of this job (my unpopular opinion)

The mainstream take is that AI is making management obsolete. I believe close to the opposite.

AI makes the good excellent and the bad worse. It amplifies whatever is already there — judgment, clarity, and direction get multiplied; so do confusion, misalignment, and bad incentives. In that world, people need more support from their managers, not less, to stay pointed in the right direction and to grow rather than just go faster.

What’s changing is how we support. I lean more on concrete, tactical guidance and less on pure coaching than I used to, because momentum matters and people often need a real next step. But the core bet is unchanged and, if anything, stronger: the human work of building fair systems, holding direction, and growing people is more valuable now, not less.

Personality quirks

• I write a lot and I prefer writing to speaking. Expect me to default to writing whenever I can — it’s how I think.
• I bring philosophy to work, increasingly a Stoic flavor of it. I’ll make odd connections and want to have the meta-conversation. Tell me if it’s distracting; technology’s effect on how we live makes it hard for me not to go there.
• I care a lot about people volunteering — speaking up, taking responsibility — precisely because I try not to dictate. Indifference is the thing I struggle most to tolerate.
• I appreciate discipline, but I won’t moralize about it or measure it. I’d rather help you find your own.

My interests

• I read a lot of SciFi and fantasy, in bursts, often finishing a book over a weekend. (Lately: the Bobiverse series.)
• I write — it’s the thing I most want to do more of, and one day, fully. I write about almost everything, which is both the joy and the problem.
• I keep a running set of Stoic reflections for engineering leaders at EM Meditations — not lessons, but a mirror.
• The rest of my time goes to my family.

And that’s me — for now. This is, as ever, a work in progress.

Written August 18, 2021 · Rewritten June 2026