Telegram RegisterThe public register of Telegram

Channel

Anton Alekseev | Инфраструктура для AI и ML

@mlops_infra

On this record: Growth · Engagement · What this channel posts · Reactions · Posts · Citations · Cite this entry

1,045subscribers

+0 since we began measuring on 7 August 2026

Risers and fallers across the register · movement among entries of 1,000–3,162.

Register entry

Telegram ID-1001871266271
TypeChannel
Username@mlops_infra
DescriptionПривет, я Антон @antonaleks Пишу здесь про инфраструктуру для AI/ML
CreatedBetween 1 October 2022 and 30 September 2023— estimated from Telegram’s id allocation, not measured. How this range is calculated.
First recorded8 August 2026
Last confirmed live10 August 2026
Measurements held2
On Telegramt.me/mlops_infra

Growth

1,0457 Aug 2026, 15:44 — 1,045 subscribers8 Aug 2026, 00:19 — 1,045 subscribers7 Aug 2026, 15:448 Aug 2026, 00:19
2 measurements taken within a single day. Dots are measurements; the straight line between them is drawn to join them, not to claim we know the path taken in between — snapshots are recorded only when a count changes, so gaps mean “no change observed”, never “interpolated”. The vertical axis spans 1,044–1,046 and does not start at zero.
Measurement log — every subscribers count we have recorded
Measured (UTC)SubscribersChange
8 Aug 2026, 00:191,045no change
7 Aug 2026, 15:441,045first reading

Engagement

20 posts held, back to 17 February 2026the reader has not yet reached the start of this channel’s public history, so older posts may sit further back, unread. Read across 1 pageof Telegram’s post history, 20 posts per page.

Nothing published in the last 30 days. ERR and ER are rolling 30-day measures, so there is nothing to compute — we hold 20 posts for this entry, the most recent from 13 July 2026. An engagement rate over an empty window would be a number about nothing.

What this channel posts

Photos
40
Videos
2
Links
82

Lifetime counters from Telegram’s own channel header, read 8 August 2026 — not the date at the top of this page, which is when the subscriber count was last read. Below Telegram’s rounding threshold, so these counts are exact.

Reaction mix

303 reactions across 19 posts, in 5 distinct kinds. The most used accounts for 42.2% of them.

Every reaction kind recorded on the sample, most used first
ReactionCountShareShare, drawn
🔥12842.2%
👍9832.3%
6019.8%
😎92.97%
👾82.64%

No sentiment is inferred, and none should be read in. This table is ordered by count and by nothing else. Emoji do not carry stable meaning across languages or communities — 🙏 is thanks in one channel and mourning in another — so we publish which ones were pressed and how often, and pass no judgement on what an audience meant by them.

Precision. Telegram publishes reaction counts per emoji and short-forms each one — 4.34K, 1.2M — so any single kind at or above 1,000 reaches us at three significant figures, and only counts below 1,000 are exact. The shares above are ratios of those figures and carry the same error. This is also why the total here can differ slightly from a reaction total printed elsewhere on the page: both are sums of the same rounded parts, taken over samples with different edges.

Coverage. Reactions were read on 19 of the 20 sampled posts in this sample. Summed by Telegram’s own count on each post — not by adding up the per-emoji breakdown above — those same posts carry 303reactions in total: the kind of figure the paragraph above means by “a reaction total printed elsewhere on the page”.

Measured over the 20 most recent posts we hold, published 17 February 2026 to 13 July 2026, using the newest reading held for each. Telegram Stars are excluded: they are a payment, not a reaction, and they have their own section.

Recent posts

13 Jul 2026, 08:51 UTC703 views34 reactionsread 8 August 2026
Photo

Ну что, цель достигнута Всем аригато за активность) Посмотрим что там за бугром творится на конференциях 🤣

🔥22😎93

8 Jul 2026, 10:12 UTC984 views13 reactionsread 8 August 2026

Друзья, моя первая папочка 📂 в телеге, да еще и с такими классными каналами про инфраструктуру для ML и LLM, а также про агентский кодинг и GPU. Актуальные ресерчи, практика на k8s, бенчи с новых моделях (и не просто репосты и ссылки на существующие, а прогоны на личных домашних мини датацентрах) Коротко по остальным ребятам: - Александр - пишет про агентский Кодинг, продукты, Go и инженерную практику, еще и со ст

👍75🔥1

7 Jul 2026, 09:11 UTC891 views14 reactionsread 8 August 2026

Всем привет! У моего товарища будет воркшоп про AI-агентов в разработке - будет полезно всем, кто хочет не утонуть в нейрослопе, который потом самому же поддерживать. Сам тоже загляну, для меня актуальная проблема вычищать нейрослоп из агентов, но времени не хватает постоянно новые подходы самому ресерчить - а тут за часик все расскажут, покажут) Будут разбирать не просто вайбкодинг, а как этим нормально управлять:

🔥9👍32

29 Jun 2026, 08:21 UTC≈1,060 views12 reactionsread 8 August 2026

ML для больших компаний. Вышла моя статья про ML-платформу на Kubeflow. Это текстовый пересказ доклада на infra.conf2026, так что кто не успел посмотреть - теперь можно ознакомиться и в текстовом виде) Также там удобно расставил ссылки на источники, например какие PR мы делали в community, где взять манифесты чтобы воспроизвести у себя платформу и другие полезные материалы

🔥9👍3

23 Jun 2026, 08:09 UTC≈1,050 views1 reactionsread 8 August 2026

#mlinfra_digest Снова подборочка от моего разнорабочего. 1️⃣ Speculation Is All You Need 🔎 О чём: Modal показывает, что speculative decoding даёт 2–3x прирост интерактивного inference и выпускает новые DFlash draft models для Qwen. 2️⃣ GKE Inference Gateway prefix caching accelerates AI inference 🔎 О чём: Google описывает prefix-cache-aware routing в GKE Inference Gateway и приводит бенчмарки по throughput, wait t

1

17 Jun 2026, 07:49 UTC≈1,380 views6 reactionsread 8 August 2026

#mlinfra_digest Всем привет! Openclaw принес мне свежую порцию интересных статей про инфраструктуру платформ, выделил наиболее актуальные. 1️⃣ Cloud native is now AI-native: Engineering production-ready AI Cloud native трансформируется в AI native. На первый план выходят отдельные primitives: DRA, Pod Groups / Workload API, Inference Gateway, плюс разговор уже идёт про AI Conformance и security для agentic flows.

🔥4👍2

11 Jun 2026, 10:12 UTC≈1,210 views11 reactionsread 8 August 2026

GPU-шеринг как у Яндекса, но на open source: Kueue Яндекс недавно рассказал про Dev Cluster, и также в каналах подозрительно много выложили постов про это (например раз и два) — самописную систему шеринга GPU между ML-инженерами: бери карту, когда нужна, отдавай, когда нет, простаивающее железо достаётся другим. Решение классное, разработчики большие молодцы - но хотелось бы и у себя потрогать это руками, а не тольк

🔥91👍1

9 Jun 2026, 09:25 UTC850 views6 reactionsread 8 August 2026

Мне тут дали 3 Cowork invites в Claude на недельку, так что налетайте на промик В целом норм история для задач не связанных с кодом. Например - заполнить доки для визы в Японию)))

👍6

8 Jun 2026, 14:36 UTC881 views61 reactionsread 8 August 2026

Конишуа ребята! Ну что, попробуем с вами немного кликбейтную историю. 50 лайков на пост и если набираем к концу июля 1000 подписчиков - лечу в Японию на KubeConf делать обзор на зарубежные конференции. Узнаем че там с DRA, инференсом в k8s, как контрибьютят в kubeflow и kserve непосредственно от контрибьютеров)

👍52👾72

4 Jun 2026, 15:41 UTC≈1,150 views18 reactionsread 8 August 2026
File

Спасибо всем, кто послушал доклад 🙌 Как и обещал, оставляю ссылки на наши proposal / PR / issue в Kubeflow Pipelines. Пулреквесты делал мой коллега Антон Печенин — Антон, тебе большое спасибо! • proposal for central driver implementation — proposal по переводу KFP с driver pod на более лёгкую central driver архитектуру, чтобы уменьшить overhead запуска задач. • central driver implementation — практическая реализаци

8👍6🔥4

25 May 2026, 13:12 UTC≈1,070 views19 reactionsread 8 August 2026
Photo

Доклад на infra.conf 2026! Привет! 4 июня в Москве на infra.conf 2026 буду рассказывать как в avito строим ML платформу на базе kubeflow. Режиссерская расширенная версия моего доклада про ml платформы. Расскажу подробно технические реализации всех компонент • очереди задач на Kueue • что мы патчили в пайплайнах kubeflow и сократили в 2 раза использование ресурсов k8s • как настроили observability и добавили интегра

17👍1🔥1

22 Apr 2026, 13:03 UTC≈1,500 views16 reactionsread 8 August 2026
File

Собрал ссылки по темам из моего доклада, чтобы можно было не только посмотреть презентацию, но и пойти дальше в детали. А также запись выступления. 1. Три подхода к построению MLOps-платформы Когда мы проектировали курс по MLOps для Яндекс Практикума, отдельно разобрали три варианта архитектуры MLOps-платформ на примерах известных западных компаний. Если хочется посмотреть на разные платформенные подходы не абстрак

🔥151

Showing the 12 most recent of 20 posts we hold for @mlops_infra. View and reaction counts are the latest single reading for each post, not a live figure, and a recent post is still accumulating both. A view count marked was rounded by Telegram before we ever saw it — t.me prints views in full below 1,000 and to three significant figures above, so ≈1,200,000 means somewhere between 1,150,000 and 1,249,999. Unmarked counts are exact. Text is reproduced from the public post preview and truncated for length.

Citation-graph rank

Citation-graph rank — 957,782 of 1,160,990entries in the measured graph. A weighted position computed from the forward and mention edges below — republished posts weigh more than named mentions — and recomputed periodically, over the whole graph. Published only as this ordinal position, never as a score: a position is a fact, and a score printed beside one channel’s name would read as a verdict this register does not make. The two counts beneath stay separate for the same reason mentions are never summed with forwards anywhere else on this page — a named-by count costs nothing to manufacture. The top 100 by this measure, or how it is computed.

Forward network

Republished by

Channels on the register that have forwarded this channel's posts into their own feed.

Built only from forwarded posts we have actually read, on both sides. Coverage is early and deliberately incomplete: a missing link means we have not read the post that would prove it, never that the relationship does not exist. Counts are distinct forwarded posts observed, so they only ever go up as we read more.

Mentions

Named by 1 registered channel — every channel on the register whose own posts have named this one, by its current username or any other username it currently holds, merged from two separately captured readings of the same fact so a namer caught by only one of them is not missed and a namer both caught is not counted twice. A username this channel has since dropped is not matched — that handle may belong to someone else now, and crediting today’s namer to yesterday’s owner would misattribute it.

Named by

Channels on the register whose posts name this channel's handle.

A mention is a weaker signal than a forward and is counted separately for that reason — naming a channel is not republishing it, and a handle in a post body is easy to place deliberately. The post counts beside each row below are distinct posts in which the handle appeared, from posts we have read on both sides — the “Named by N registered channels” figure above is a different count, of distinct NAMING CHANNELS rather than posts, and is not the sum of the rows under it.

Cite this entry

A live page changes as we take new readings, so a citation should name the measurement it is based on, not just the URL. The line below cites the subscriber count as measured 8 August 2026 — this entry's latest reading, not the date you are reading this.

“Anton Alekseev | Инфраструктура для AI и ML” (@mlops_infra), 1,045 subscribers as measured 8 August 2026. Telegram Register, tgregister.com/channel/mlops_infra.

Full measurement history, CC BY 4.0. Every reading this register holds for this entry, not just the latest one, as a dated, downloadable record: CSV · JSON. Free to use with attribution to tgregister.com. Each file carries its own generation timestamp, which is the figure to cite for exactly when the data was retrieved.