Changes I made after the previous post about my job-vacancies channel

✏️ I want to talk about what I eventually changed following my previous post about my vacancies channel.

🔸 On the same day I published that post, I slightly rewrote the bot so that vacancies would arrive in my personal messages instead of directly going to the channel, with approval and delete buttons. So for the last three days I have been manually selecting the five best vacancies. In the future I’m considering delegating this to a neural network, or configuring better filters.

🔸 Today I completely reorganized the architecture of the project. The goal was to avoid overloading the system with too many files, even while adding a new service, two payment systems, and integrating a PostgreSQL database into the project.

Over the next few days I will be testing the system before presenting it in the channel. You can also take a look and try it here: @publisher_private_bot

It is expected that the public vacancies channel will now contain only a few of the best vacancies so that it doesn’t become overloaded, but anyone who wants can continue receiving all vacancies in their private messages. Users will also be able to set their own filters for vacancies and exclusion keywords.

💫 The first two weeks are a trial period. At any moment the user can purchase one month of access or a subscription, and it will be added on top of the remaining trial period.

The positioning of the channel will gradually shift from an aggregator channel to a curated vacancies channel where companies can purchase placement for vacancies or resumes, plus a vacancy distribution service for job seekers.

Updates to the women’s cycle tracking bot

Hello dear ones 🧡 Update on the women’s cycle bot @RitmaOn_bot.

✨ Since my last post about this project, there have been a couple of important changes in how it works:

1️⃣🔸 Added the ability to record events retroactively using a calendar.
This includes events such as the start of menstruation, sex, and symptoms. What does this mean for the user? If you forgot or couldn’t record an event at the moment it happened, you can now add it later on the correct date. Additionally, when starting the bot you can immediately enter the dates of your last three periods. This allows the bot to calculate all cycle phases right away and notify both you and your partner.

2️⃣🔸 A bug has been fixed.
Situation:
After New Year I started receiving phase notifications every day, even though they should only be sent once at the beginning of each phase. It took me a while to understand what was happening because technically I had already solved this case before, and the bug appeared only for me — other users did not experience it. Eventually I discovered that my ex had blocked the bot, so the database stopped recording that the notification had already been sent.

Solution:
Now the notification is first sent to the main user and recorded in the database as delivered, and only after that it is sent to the partner, where the status no longer affects the logic.

3️⃣🔸 Based on the previous issue, I added an important new feature:
a button to remove the partner and automatically generate a new invitation link, so a user can add a new partner or simply continue without one.

✨ Upcoming updates:
I see that many of you want the ability to add multiple partners. This will be implemented in one of the future updates. Feel free to message me if you have additional feature requests :)

Thinking about monetization of my gamedev job channels

✏️ Today I sat down to think about monetization of my gamedev job channels.

⭐️ https://t.me/works_gamedev — 3300 members
⭐️ https://vk.com/works_gamedev — 6350 members

Something strange happened: when both channels doubled in size, the number of paid advertising orders almost dropped to zero. Before that there were consistently 5–6 placements per month. What makes it even more surprising is that previously most of the audience consisted of bots, while now the channels have a much more relevant live audience, views have grown significantly, and the channels are fairly well known in the Russian-speaking gamedev market. Information about paid placement (and the price) is clearly visible in the channel.

I am currently trying to understand the reasons behind these changes. One possible explanation is that many people clearly see the channel as an aggregator, so they assume their vacancy will appear there for free anyway. But then the question remains: why are resume placements being purchased less often, if resumes are not posted for free at all?

One idea is to limit posting to the five best vacancies per day. This would create scarcity of free posting slots in the channel. At the same time it could increase the number of views per post, since attention would no longer be spread across twenty different posts.

Another idea is to introduce a personal subscription through the bot, so that people who want to receive every vacancy can get them directly in private messages as soon as they appear.

⛵️ Still thinking about it. I will gradually introduce small changes and observe how they affect the channel, user activity, and revenue.

Working on my LinkedIn SSI and repositioning my profile

🔥 Hi everyone.

You know that I’ve been actively working on my LinkedIn. (I once found a job there and now I can’t stop.) Not long before the New Year I completely changed the positioning of my profile, because I decided to finish my career in gamedev and fully move into Python parsers, automation systems, and bots.

Over the past few weeks I have been working on my SSI. This is a trust metric from the LinkedIn algorithm that affects whether the platform promotes your profile and your posts. You can check your own score here: https://www.linkedin.com/sales/ssi

As part of this, I started publishing technical posts once per week, talking about my personal projects and test tasks that I completed. I also began adding people from the industry I’m moving into to my network.

Over four weeks this resulted in:

1️⃣ SSI growth from 46% to 50%
2️⃣ The speed at which my posts accumulate views increased slightly but noticeably
3️⃣ I received two interview invitations in my private messages from hiring managers during this period
4️⃣ After my latest post today (https://www.linkedin.com/posts/helena-gusarevich_%D0%BA%D0%B0%D0%BA-%D1%8F-%D0%BE%D0%B1%D0%BE%D1%88%D0%BB%D0%B0-%D0%B7%D0%B0%D1%89%D0%B8%D1%82%D1%83-%D0%BE%D1%82-%D0%B1%D0%BE%D1%82%D0%BE%D0%B2-%D0%B8-%D0%BD%D0%B0%D1%81%D1%82%D1%80%D0%BE%D0%B8%D0%BB%D0%B0-activity-7413066583268495360-J1WU), for the first time I received relevant technical comments from specialists under the post. My recommendation feed also started showing many technical posts, which means that if I see posts from my niche, people from my niche can also see my content. The algorithms seem to have calibrated, and I should start appearing in search results when people look for specialists with my profile.

⭐️ There won’t be a conclusion here. But I highly recommend paying attention to this metric when developing your account.

Air quality desktop widget for Windows

🙊 🔥 If you follow news from Armenia, you probably know how bad the air in Yerevan gets during winter heating season. This year it has broken all previous records.

✨ Of course I bought a respirator and a home air purifier, but you never know where inspiration will catch you and make you want to code something.

🐈 So last week I built myself a desktop widget for Windows. I’m extremely happy with it and keep looking at it during the day thinking how cool it turned out.

❤️ How it works: configuration through .ini files, rendering with Shape meters, external data via Python scripts, and the interface itself rendered with Rainmeter.

❤️ The script fetches data directly from nearby public air-quality sensors (within about one kilometer) and passes it to the interface, while also generating graphs. Data updates every five minutes.

The arrows (which I drew myself ^^) show whether the indicator is going up or down. The color of the circles reflects pollution level according to the scale. When you click on the circles, individual pollution graphs for the past day open, along with the current particle concentration values in the air with the same color markers.

Rainmeter settings allow me to control how the widget behaves: edge snapping, transparency, window layer priority, and whether the widget is clickable.

❤️ I’ve also started posting some of my projects on LinkedIn, because I’m leaving gamedev and shifting my profile toward Python engineering. If you want to support it, you can like the post here .

Major update for the menstrual cycle Telegram bot

🔥 I updated my Telegram bot for menstrual cycle tracking @RitmaOn_bot.

⭐️ Registration and partner linking

• The main user (a woman) registers through /start.
• She receives a unique invitation code for her partner.
• The partner registers using that link and becomes connected to the user.
• The partner only sees information about the start of new cycle phases and the moment when the user marks the beginning of a new cycle. All other data remains private for the main user.

⭐️ Main menu features (/menu)

• Mark the start of menstruation for a specific date.
• Record sexual activity for the current day with options “protected” or “unprotected”.
• Add symptoms, feelings, and wellbeing notes in text format.
• View a menstrual calendar for the past year.

⭐️ Cycle calculation and prediction

• Automatic calculation of the average menstrual cycle length.
• Detection of the current cycle phase: menstruation, follicular phase, ovulation, luteal phase, and premenstrual phase.

⭐️ Notifications

• The bot tracks cycle phases automatically and sends notifications once per phase.
• Notifications are delivered both to the user and the partner.
• Messages include useful recommendations for each stage of the cycle, tailored separately for the user and the partner.
• Warnings about upcoming menstruation one day before it begins.
• Notifications about delays with recommendations to take a pregnancy test or consult a doctor.

⭐️ Reports for a gynecologist (/report)

• An extended report covering the last 6 months, including:
• Average cycle length
• Detailed list of recent cycles
• Symptoms recorded over the last two weeks
• Statistics of sexual activity over the past three months with dates and protection status

⭐️ Technology stack

• Telegram Bot API with python-telegram-bot for Telegram integration
• MongoDB running in Docker for storing users, events, and notifications
• Daily data updates and cycle calculations handled by apscheduler
• Infrastructure running on a Linux VPS

The architecture allows the bot to scale and support additional features and new users in the future.

🔥 See you in the next posts 🙂 🧡

LangGraph TimeBot test task

🔥 Hi!

🍂 Today I worked on a test task for this vacancy .

🍂 The full description of the test task is available here .

🧡 Result:

The final project is available on GitHub here .

The project fully meets the task requirements: it implements a stateless chatbot built with LangGraph, using a tool get_current_time that returns the current UTC time in ISO-8601 format.

The bot works both in console mode via python main.py and through langgraph dev (tested in LangGraph Studio). All setup instructions, dependencies, and screenshots of the bot running (both in console and in Studio) are included in the README.md.

Improving the snake rendering logic

🔥 I spent the entire day drawing snake segments in Gimp and configuring the rendering logic. (The first screenshot was made in Figma.)

🍁 The second screenshot shows how it used to look on my website, and the third one shows how it looks now.

The snake now opens its mouth when it gets close to the apple 😊

You can try the game yourself (from the desktop version of the site) here .

Reworking the architecture of my gamedev vacancies channel

😏 Good news — we can now see that my updated posting system for the vacancies channel is working. There were no vacancies in the Russian-speaking gamedev space for two days, and at first I thought my parsing system had broken.

💫 I learned that 👍👎 buttons increase audience engagement, which is exactly what I need in order to increase revenue from the channel.

Right now the channel has about 1500 subscribers. A sponsored post costs 1000 RUB and is purchased around 2–3 times per month, usually for the same vacancy or resume.

✳️ The main challenge is that Telegram bots cannot read channels where they are not admins. My system parses vacancies from 15 other channels across Telegram, because the original idea of the project was simple: users shouldn’t need to search across the entire platform anymore. So historically the vacancies were collected and published by a userbot.

✳️ Unfortunately, a user account cannot create buttons under posts — only the Telegram Bot API can do that. Because of this, the system now uses two bots:

• The first bot scans Telegram channels for vacancies and stores the text and images in a small internal database.
• The second bot checks for new entries in the database, publishes them with engagement buttons, and then removes them from the queue.

✳️ In addition to like/dislike buttons, I also added buttons for submitting a resume to my candidate database (another project of mine), as well as a button allowing users to subscribe to that database.

🥰 For now I hope this won’t look like overly aggressive promotion and will instead help grow the candidate database faster while also increasing subscription traffic.

A day of personal projects and experiments

🧡 Yesterday was my first day without client orders, so I decided to take a break and work on whatever came to mind — mostly my own small projects.

💕 What happened during these two days:

✳️ I deployed and configured a copy of one of my parsers for my daughter Venus. It now posts Fortnite news into her channel where she occasionally shares her streams: @venera_i_hel.

✳️ I updated my vacancies bot @works_gamedev. The architecture became significantly more complex. Originally it was a userbot, but I learned that like/dislike buttons dramatically increase engagement in Telegram channels. Since userbots cannot attach buttons to posts, I ended up rebuilding the system so that multiple bots now cooperate. After successful testing I plan to write a separate post about the architecture.

✳️ I added another good meme group to my meme parser, so the content flow in my meme channel is now essentially endless: @norm_mem.

✳️ And today I built a Telegram bot for menstrual cycle tracking: @RitmaOn_bot. This one is honestly my favorite project right now.

📖 Background: I always hated how apps from marketplaces treat a missed entry of the start of menstruation as a delay. Immediately the internet fills with baby product ads. I always considered this kind of health-data leakage too invasive to tolerate, so I decided to build my own solution.

What the bot does:

• The main user registers as a woman and receives an invitation link for a partner. A person registering through that link receives the partner status.
• Through the /menu command the user can:
  • mark the start of menstruation
  • record sexual activity for the day (protected or unprotected)
  • log symptoms or unusual feelings
  • export cycle and health data for the last three months before visiting a gynecologist
• The bot calculates the average cycle length and predicts ovulation automatically.
• Key feature: the bot sends notifications at the beginning of each cycle phase. The user receives information about what to expect, while the partner receives recommendations about how to interact with the user during that phase.
• Because the bot has its own database, new users can easily connect to the system in the future.

👻 Tech stack:
Language: Python
Libraries: python-telegram-bot, pymongo, apscheduler, python-dotenv
Database: MongoDB (Docker)
Hosting: Linux VPS
API: Telegram Bot API