KalmykovIT— k1 UTC
[01]Services[02]Numbers[03]Cases[04]Blog[05]About[06]Contact
Portrait of an engineer, rendered as ASCII character art
build it right_
K. Kalmykov — Sheet 1
Instrument for building software.
kalmykov — k1
uptime ...... 8h 44m 00s
cpu ......... ██░░░░░░░░░░ 18%
mem ......... █████░░░░░░░ 42%
deploys ..... 14 today
engineers ... 12 online

Numbers.
People.
Machines.

We design, build and run software systems — from the first audit to stable operation in production.

f(x) = Σ wᵢxᵢ + b·λ = 0.044·p99 < 120 ms·∂L/∂w → 0·uptime = 99.95%·n = people × method × time·O(n log n)·σ² = E[(X − μ)²]·deploys / day = 14·H(X) = −Σ p log p·commit 7f3a9c1·f(x) = Σ wᵢxᵢ + b·λ = 0.044·p99 < 120 ms·∂L/∂w → 0·uptime = 99.95%·n = people × method × time·O(n log n)·σ² = E[(X − μ)²]·deploys / day = 14·H(X) = −Σ p log p·commit 7f3a9c1·
Fig. 1 — 4 · What we do

Four instruments

Each figure is a discipline we run end to end.

Fig. 1001.k1
Software — illustrative rendering

Software

Web platforms, internal systems and APIs built around business needs.

TypeScriptGo
Fig. 2002.k1
Data & ML — illustrative rendering

Data & ML

Data pipelines, forecasting models and analytics people actually use.

PythonML
Fig. 3003.k1
Interfaces — illustrative rendering

Interfaces

Products shaped around the people who operate them every day.

DesignReact
Fig. 4004.k1
Infrastructure — illustrative rendering

Infrastructure

Cloud, CI/CD, monitoring and on-call after launch.

KubernetesSLA
Readout · numbers we report against
n = people × method × time
99.95%
availability across systems we operate under SLA
MetricLevelValue
Release cadence██████████████░░14 / day
Test coverage█████████████░░░78%
Incidents / quarter██░░░░░░░░░░░░░░1.2
Avg. response██████████░░░░░░< 4 h
Process list · selected work

Cases

All cases →
PIDNameClientStackStatusYear
0001Smart Notes SkillOpen sourceClaude Skill · Markdown · Obsidianopen source · MIT2026
Profile · who is behind it

About

I'm Mykhailo Kalmykov, founder of Kalmykov IT.

The world is shifting at a decisive point: you no longer pay for time, you pay for results. For decades, service meant someone sitting X hours on a task, and you paid for those hours regardless of outcome. AI-driven automation flips that: a system does the task, reliably and without a timesheet. You stop paying for presence and start paying for something working.

That's exactly what I build for small and mid-sized companies: systems that don't just do routine work — invoices, scheduling, customer requests — faster, but on their own.

get in touch.

Input · start a conversation

Tell us the problem

We reply within one business day; the first meeting is free.

kalmykov — contact — 80×24tty1
$ kalmykov --contact
your name:
enter ↵ to continue