Posted without readable text
❤17🔥6🦄4❤🔥1👍1

Channel
@yandexforml
On this record: Topic · Growth · Engagement · What this channel posts · Reactions · Posts · Posts edited after publishing · Citations · Handles named that no longer answer · Telegram's recommendations · Cite this entry
19,153subscribers
+510 since we began measuring on 6 August 2026
Risers and fallers across the register · movement among entries of 10,000–31,623.
| Telegram ID | -1001433806174 |
|---|---|
| Type | Channel |
| Username | @yandexforml |
| Description | 💙Канал для ML-сообщества от Яндекса и место встречи специалистов в сфере анализа данных. Чат→ https://t.me/+XR1fd2QNnaIxNTUy Вопросы: @Ekaterina_Lyagina Все каналы по стекам: https://t.me/addlist/Hrq31w2p1vUyOGZi |
| Created | Between 1 April 2019 and 31 October 2021 — estimated from Telegram’s id allocation, not measured. How this range is calculated. |
| First recorded | 6 August 2026 |
| Last confirmed live | 15 November 2026 |
| Measurements held | 61 |
| Confirmed unchanged | 1 time, most recently 15 November 2026 |
| On Telegram | t.me/yandexforml |
Technology — a classification, not a measurement. An on-box language model (Qwen3.6-35B-A3B-FP8, prompt version 1) read this channel’s own recent posts on 10 September 2026 and assigned it the closest of 31 fixed categories, at 88% confidence. This is a model’s judgement about what the channel is likely to be about, not a fact this register measured the way a subscriber count or a view count is measured — it can be revised on a later pass, and it carries no weight anywhere else on this page. How this classification works, and the other topics.
| Measured (UTC) | Subscribers | Change |
|---|---|---|
| 15 Nov 2026, 02:59 | 19,153 | -2 |
| 11 Nov 2026, 21:00 | 19,155 | +19 |
| 9 Nov 2026, 20:41 | 19,136 | +10 |
| 8 Nov 2026, 11:18 | 19,126 | -6 |
| 7 Nov 2026, 03:57 | 19,132 | +2 |
| 6 Nov 2026, 03:41 | 19,130 | +1 |
| 5 Nov 2026, 06:10 | 19,129 | +4 |
| 4 Nov 2026, 07:51 | 19,125 | +1 |
| 3 Nov 2026, 10:38 | 19,124 | +1 |
| 1 Nov 2026, 05:27 | 19,123 | -1 |
| 31 Oct 2026, 05:48 | 19,124 | +5 |
| 30 Oct 2026, 04:45 | 19,119 | +2 |
| 29 Oct 2026, 03:35 | 19,117 | +5 |
| 27 Oct 2026, 23:50 | 19,112 | +7 |
| 27 Oct 2026, 01:10 | 19,105 | +6 |
| 26 Oct 2026, 01:47 | 19,099 | +8 |
| 25 Oct 2026, 03:46 | 19,091 | +4 |
| 22 Oct 2026, 01:16 | 19,087 | +14 |
| 20 Oct 2026, 22:19 | 19,073 | +7 |
| 20 Oct 2026, 00:28 | 19,066 | first reading |
32 posts held, back to 30 July 2026 — the reader has not yet reached the start of this channel’s public history, so older posts may sit further back, unread. Read across 84 pages of Telegram’s post history, 20 posts per page.
ERR is average views per post over the last 30 days divided by subscribers, the definition TGStat uses, so this figure is comparable with the one you will see elsewhere. It falls structurally as a channel grows: a high ERR on a small channel and a low one on a large channel describe reach mathematics, not quality. We publish the figure and the sample it came from and pass no verdict on it.
ER is defined industry-wide as (forwards + reactions + comments) ÷ views — note the denominator is views, not subscribers. Telegram’s public web preview carries views and reactions but not forward or comment counts, so the reaction rate above is the reactions term only and is therefore a floor: the true ER for this channel is higher by an amount we have not measured and will not estimate.
| Window | Rolling 30 days · latest post in window 13 November 2026 |
|---|---|
| Posts held | 32 (30 July 2026 – 13 November 2026) |
| Views total | 14,620 |
| Reactions total | 127 |
| Forwards / comments | not exposed by the public surface — not measured, not estimated |
| Readings taken | 17 Nov 2026, 05:05 UTC |
Views are a single reading per post, taken at the time above. A post published in the last day or two is still accumulating views, which pulls the 30-day average down slightly. That is a property of the standard definition rather than a fault in it, so we keep the definition rather than “correcting” the number into something nobody can reproduce.
Precision. Telegram publishes view counts on its public widget in short form — 8.12K, 3.7M — so any reading at or above 1,000 reaches us rounded to three significant figures, and only counts below 1,000 are exact. Averages and rates derived from them are shown to the same precision rather than to the unit: a figure like 3,701,250 would assert digits nobody measured.
Reaction counts are published per emoji and rounded the same way, so a total below 1,000 is exact and a larger one is a sum that may carry a rounded component from each emoji above 1,000. Because it is a sum, it does not look rounded — read a large reaction total as three significant figures per contributing emoji rather than as the figure it prints.
Lifetime counters from Telegram’s own channel header, read 17 November 2026 — not the date at the top of this page, which is when the subscriber count was last read. A count marked ≈ was rounded by Telegram before we ever saw it — t.me prints these counters in full below 1,000 and to three significant figures above, so ≈142,000 means somewhere between 141,500 and 142,499.
Measured directly from 3 videos with a duration reading, out of the posts we hold for this channel — not this channel’s whole posting history, only the sample this register has actually read. An exact reading to the second, taken from the post itself rather than from Telegram’s own rounded chrome, so it carries no ≈ mark.
915 reactions across 31 posts, in 16 distinct kinds. The most used accounts for 45.6% of them.
| Reaction | Count | Share | Share, drawn |
|---|---|---|---|
| ❤ | 417 | 45.6% | |
| 🔥 | 210 | 23.0% | |
| 👍 | 77 | 8.42% | |
| ❤🔥 | 65 | 7.10% | |
| 👏 | 47 | 5.14% | |
| 👎 | 23 | 2.51% | |
| 🎉 | 20 | 2.19% | |
| 🦄 | 18 | 1.97% | |
| 😱 | 11 | 1.20% | |
| 👀 | 9 | 0.984% | |
| 😁 | 6 | 0.656% | |
| 🥰 | 4 | 0.437% | |
| 😍 | 3 | 0.328% | |
| ✍ | 2 | 0.219% | |
| 😢 | 2 | 0.219% | |
| 🤩 | 1 | 0.109% |
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 31 of the 32 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 915 reactions in total: the kind of figure the paragraph above means by “a reaction total printed elsewhere on the page”.
Measured over the 32 most recent posts we hold, published 30 July 2026 to 13 November 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.
Telegram’s own ‘edited’ marker: 0 of 4 posts held from the last 30 days (0%) carried it when this register last read them. The marker says a post was changed after publishing, not what it said before — this register holds the earlier wording only where an edit fell between two of its own reads (see the edit ledger and how it is measured).
Posted without readable text
❤17🔥6🦄4❤🔥1👍1
🦠 От органической химии к исследованиям AI Что делает исследование по-настоящему качественным и какие задачи сегодня стоят перед ML-индустрией? 👷 Об этом рассказывает Сергей Кастрюлин, Head of Applied Research в Yandex Research. В карточках собрали его мысли о пути в ML, работе исследователей и будущем AI. 🔳 Кстати, у коллег из Yandex Research появился собственный телеграм-канал, где команда делится научными стать…
❤21🎉12❤🔥7👍2
🧿 RuntimeNodes: как ML-разработчики в Яндексе начали писать рантайм сами Привет, это Ильнур Хузиев, руководитель службы генеративного ответа Алисы AI. Я работаю в команде, которая переносит ML-эксперименты для генеративных ответов Алисы AI в Поиске из офлайна в продакшен. Офлайн‑прототипы у нас обычно строятся поверх MapReduce‑хранилища. В Яндексе для этого используется YTsaurus. Пайплайны, как правило, описываются…
❤12👍2🔥2😁1
📺 Practical ML Conf: записи всех докладов Смотрите выступления про лучшие ML-практики и вдохновляйтесь. Кто знает, куда вас заведут эксперименты: может, вы станете одним из спикеров конференции 2027 года. Записи докладов мы разделили по залам: ⚪️ «Данные» на ютубе и в VK Видео ⚪️ «Код» на ютубе и в VK Видео ⚪️ «Сеть» на ютубе и в VK Видео ⚪️ «Серверная» на ютубе и в VK Видео Keynotes выделили отдельно: ⚪️ Доклад…
❤28🔥8😍3
🪧 Строим логистику пути в ML Привет всем, меня зовут Фархад Сабилов, я руковожу группой ML игровых и коммуникационных механик в Яндекс Плюс AdTech. Ещё во время учёбы в университете я устроился на первую фултайм-работу: начал помощником логиста и за три года стал руководителем сразу в трёх компаниях холдинга. В какой-то момент мне захотелось сменить карьерный трек. Несколько лет я искал себя, попробовал открыть сво…
❤11🔥5👍3👀2
🔎 Открываем претрейн Alice AI Search Быстрый ответ Алисы AI — это самый массовый генеративный продукт Яндекса и первая точка контакта с Алисой для пользователей Поиска. Даже в час пиковой нагрузки он должен появляться за секунды. Поэтому мы, команда Alice AI Search, решили адаптировать весь пайплайн: от собственного претрейна с кастомной архитектурой до online RL на поведенческих сигналах аудитории. Alice AI Search…
❤12👏4✍1
📕 Выкатываем AliceAI-Foundation-80B-A3B-Base в опенсорс, занимаемся ML-разработкой без написания кода и (внезапно) пишем рантайм. Об этом и не только — в новом ML-дайджесте 🚕 Куда сходить 🔴 6 октября — Архитектурный клуб Яндекс 360 онлайн. Открытое сообщество для архитекторов и инженеров, которые строят высоконагруженные системы. На этот раз разберём архитектуру AI-агента над корпоративными данными: от оркестратора…
❤15❤🔥5👍5🔥2😁1😢1
💻 Как заниматься ML-разработкой без написания кода 👋 Всем привет, это Руслан Кулиев, разработчик в службе клиентского ML в Яндекс Лавке. Последние полгода я развивал наши трансформерные технологии на базе Argus и локальную среду для DL-экспериментов. За это время из более чем 100 тысяч строк финального кода я не написал ни одной: разработка и эксперименты были целиком AI-first. 👷 В карточках я поделился некоторыми …
❤10🔥3👀1👍1
🧑🏫 Data Dojo: встреча ML-сообщества в Москве 🔳 28 октября мы соберёмся на Data Dojo — митапе для тех, кому интересно машинное обучение. Здесь вас ждут прикладные доклады от экспертов, нетворкинг, мастер-классы, мерч и многое другое. На ивенте начинающие ML-специалисты смогут попасть в тусовку, познакомиться с трендами индустрии и получить карьерную консультацию. А опытные инженеры — найти единомышленников и завест…
❤31🔥21👍17❤🔥3🦄1
🎉 Practical ML Conf 2026 завершилась, но впечатлений хватит на весь год 🔳 Последние новости с конференции: спикеры выступили, Yandex ML Prize вручили, подарки раздали, дискуссии провели. Получилось очень здорово: участники знакомились с сервисами Яндекса, пробовали новое, нетворкали и делились инсайтами. Спасибо всем, кто был с нами: конфу в Москве посетили 870 участников, а онлайн-трансляция собрала около 2000 зри…
❤🔥15❤9🎉7👍1🔥1
🧿 Одна модель вместо всего рекомендательного стека Яндекс Музыки Привет, это Николай Савушкин, руководитель службы рекомендательных технологий Яндекс R&D. В августе мы опубликовали Sona Technical Report — результат года нашей работы над новой генеративной end-to-end-моделью рекомендаций. Раньше все рекомендации в Музыке проходили через многоступенчатый каскад: отбором занимались больше 15 кандидатогенераторов, а за…
🔥19❤11❤🔥5✍1🦄1
🎓 Как преподаватель математики стала руководителем ML-направления в ШАДе Привет, это Валя Бронер. Я 11 лет преподавала студентам математику в Томском государственном университете, а затем перешла в бигтех. Последние три года я руковожу ML-направлением в ШАДе Яндекса и создаю курсы по машинному обучению. Сегодня я расскажу, что мне нравится в преподавании, почему решила уйти из вуза в ШАД и как ML-разработчики помог…
❤23🔥18❤🔥7😱6👎1😢1
Showing the 12 most recent of 32 posts we hold for @yandexforml. 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.
@yandexforml edited 1 post after it first published — the same permalink now carries different wording than the one this register originally read, caught because our own crawl held a copy of the earlier text.
An edit is not deception. Typo fixes, price updates and corrections look exactly like this too — this register can tell you the wording changed and when, not why. How this is measured.
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.
Named by 7 registered channels — 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.
Names
Channels on the register whose handles appear in this channel's posts.
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.
@yandexforml named 1 handle that resolve to nothing today. That is a fact about the reference, not necessarily a fact about the handle’s history — see the two groups below.
Most of these may never have existed as a live channel at all. A handle a channel names can be a typo, an aspirational name nobody registered, or a channel that was already gone before this one ever mentioned it. Unless a row below is marked evidenced, all we know is that it references a handle that is not a live channel today — not that anything “died”. How this is measured.
References a handle that is not a live channel — we have no record it ever was one.
Telegram’s own answer, not this register’s. When this register asks Telegram’s API what is similar to this channel, this is the list it returns, in the exact order Telegram returns it — never re-sorted by subscribers or by anything else this register measures. The relationship, and the order, are Telegram’s; we record them and date them, and make no claim of our own about which of these channels actually resemble this one.
Read from Telegram’s recommendation API, most recently 4 October 2026. Telegram holds a list like this for a small and growing share of the register — how this is measured, and why most channel pages show nothing here.
The reverse of the list above, and a different kind of signal. This does not require this channel to have ever been asked about directly — each row below is a channel we DID ask Telegram about, whose Telegram-generated list happened to include this one. A channel can appear here with an empty list above it, because being named by someone else’s query is independent of having been queried itself.
This channel appears in 39 seed channels' Telegram-generated recommendation lists in total, of which the 24 where it ranks highest are shown above. Each is Telegram’s list for THAT channel, not this one — see how this is measured.
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 15 November 2026 — this entry's latest reading, not the date you are reading this.
“Yandex for ML” (@yandexforml), 19,153 subscribers as measured 15 November 2026. Telegram Register, tgregister.com/channel/yandexforml.
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.