Tadeush Muratov
Offer Tadeush work on your next project.
Private message
Offer work
Dubai,
United Arab Emirates
10 days 13 hours back
available for hire
2 proposals made
age 33 years
on the service 15 days 17 hours
Rating
Successful projects
Average rating
Skills and abilities
Programming
Services
Photo, Audio & Video
Promotion
Architecture
Mobile development
Writing
Portfolio
-
500 USD RAI bot
AI & Machine Learning
Forbes Ukraine about my works)
-
20 000 USD Company OS
AI & Machine Learning
Company OS — an internal AI operating system for automating the creation and publication of digital books in Shopify. The product works like a "Book Factory": the user sets the book's topic, after which the Director launches a chain of specialized bots that conduct pricing research, gather materials, write a draft, edit the text, create the cover/PDF, and publish the result in Shopify.
The main value of the case: automation of the complete commercial cycle "idea → finished digital product → publication in the store," with a dashboard interface, launch history, artifacts, reports, key settings, and Shopify integration.
… Business logic
The pipeline is built around the Director Bot:
pricing / pricing_finance_bot — analyzes the market, competitor prices, floor price, penetration discount.
research — gathers sources and bullet points on the topic.
writer — generates the book draft.
editor — improves the structure, tone, and quality of the text.
cover — creates the cover or placeholder.
formatter — compiles Markdown/PDF.
publisher — prepares the publication and sends the product to Shopify via gateway/API.
-
20 000 USD Jarvis AI
AI & Machine Learning
Jarvis is a personal autonomous AI system built as a "mini brain" with memory, sensors, self-state, a research cycle, external actions, and a control panel. The project addresses the task of not just an ordinary chatbot, but a local/server operational assistant that can observe events, maintain long-term memory, initiate research, make decisions based on policy/guardrails, and interact with external services through controlled "digital hands."
The system consists of several layers: brain runtime, audio/event sensors, SQLite memory layer, FastAPI dashboard, real-time WebSocket UI, LLM/research pipeline, ActionGate/Police safety layer, external connectors, and business/financial modules. The project clearly defines stages of autonomy: Stage 5 External Autonomy, Stage 6 AGI Foundation/Operations, and Stage 7 Autonomous Business.
… What the product does:
- Background Brain service with an autonomous cycle, goals, interests, reflection, and learning.
- Collection of audio events through VAD, saving episodes and brain states.
- Chat with layered persona, owner profile, owner notes, dictionary, and research context.
- Web research pipeline: search, fetch pages, summary, saving brief in knowledge base.
- Dashboard V1/V2 for managing the brain, events, memory, goals, research, business, and external integrations.
- Safety contour: ActionGate, Police Department, risk limits, financial execution guard, kill switch.
- External actions: browser automation, Gmail/email, Telegram, Moltbook, CompanyOS, Binance, Freedom24, MT5, Shopify.
- AGI/Autonomy monitoring: baseline snapshots, gap review, advancement cycles, memory quality gate, auto-audit.
-
35 000 USD Whales Officer is an AI service for automated KYC.
AI & Machine Learning
Whales Officer is an AI service for automated KYC/Compliance client verification through video interviews with a live AI avatar.
The scenario is as follows:
… The client arrives via a link or through the main page with parameters name, language, country, gender, age.
Frontend creates an interview through the Next.js API route.
Backend creates a user, generates an access_token, saves the interview state in Redis, and basic information in PostgreSQL.
Frontend creates a LiveAvatar/HeyGen context with a KYC prompt.
The client engages in a conversation with the AI compliance officer, who asks mandatory KYC questions:
source of funds;
purpose of the account and expected transaction volume;
beneficial ownership / who controls the account.
During the session, the frontend records the user's video, microphone, and avatar sound, cuts the recording into chunks, and sends them to the backend.
Backend places the chunks in a Redis queue.
A worker assembles the .webm, pulls the transcript, generates a summary via OpenAI, sends the video to an external Whales API/S3 flow, and updates the call record.
Thus, the product is not just a chatbot, but a KYC video-interview pipeline: the AI avatar conducts a regulated interview, and the system saves the evidential record, transcript, and summary for further compliance processing.
-
3000 USD RAI SALES AI
Data ParsingRAI SALES AI — an internal AI system for finding, qualifying, and bringing leads to dialogue in Telegram Business/CRM.
The project addresses the task of finding potential clients from Telegram/Threads, automatically filtering out noise, writing the first careful outreach, receiving responses in Telegram Business, continuing the dialogue through an AI assistant, recording everything in CRM, and handing over warm leads to the manager.
… The system is built as a production backend that operates on a server independently of a local Mac. The main focus is not on "chat-bot for the sake of chat-bot," but on the connection: search → triage → first contact → dialogue → CRM → handoff → quality monitoring.
What it can do:
- Accepts leads from AI Clients Finder via signed webhook.
- Processes Telegram Business webhook from `RAI_sales28_bot`.
- Engages in dialogue with the client through AI Sales Agent.
- Separates the first outreach and continuation to avoid message duplication.
- Supports pauses after a real manual response from the operator.
- Does not pause after automatic outreach echo.
- Filters out irrelevant applications: job vacancies, performers, SMM/video noise, job seekers.
- Stores candidates, leads, conversations, activity logs, alerts, service heartbeats.
- Works with Threads: inbox, triage, web publish fallback, local/server discovery.
- Has an AI CMO layer: content bank, publications, QA, external accounts, metrics, blockers.
- Has an admin/CRM UI for operators.
- Has systemd-supervised runtime on VPS.
-
3500 USD RAI parser bot
AI & Machine Learning
RAI parser is an MVP operating system for Telegram lead generation and AI sales in the Gerchik/GTE ecosystem. The project covers the path from finding potential clients in Telegram groups to the first touchpoint, continuing the dialogue with an AI assistant, warming up the lead, and transferring it to Zoho CRM.
The system consists of an operator dashboard, a Telegram user account worker for searching/autojoining/first DM, a Telegram Business Bot for incoming dialogues, an AI agent on OpenAI, a knowledge base, and CRM handoff. A Threads section is separately planned, but in the current code, this is an intentional MVP placeholder, not a working collector.
… What the product can do:
- Create search topics: keywords, language, geo, Telegram/Threads channels.
- Search Telegram groups through a real Telegram user account, not Bot API.
- Autojoin found groups with a daily limit.
- Scan messages in groups and select leads based on buyer intent: keywords plus explicit requests for help/training/consultation.
- Filter out promotions, advertisements, and irrelevant messages.
- Queue the first DM from the Telegram user account.
- Conduct Telegram Business dialogues via webhook.
- Generate responses through OpenAI Responses API with fallback responses in case of errors/timeouts.
- Process text, voice messages, images with caption/fallback.
- Transcribe voice through Whisper.
- Generate voice replies through OpenAI TTS.
- Store prompts, a knowledge base up to 100 MB, Google Docs/Sheets/Slides links, and a media bank.
- Send materials from the media bank via the service marker MEDIA:.
- Identify warm leads based on interest signals, phone, email, or consent to continue in Telegram.
- Send warm leads/leads to Zoho CRM.
- Show the operator dashboard: topics, keywords, leads, dialogues, settings, logs, Zoho events, system statuses.
Full stack:
Backend:
- Python >=3.11
- FastAPI
- Uvicorn
- Starlette SessionMiddleware
- Jinja2 templates
- SQLite via standard sqlite3
- httpx for external HTTP APIs
- OpenAI Python SDK
- Telethon for Telegram user account automation
- pytest for tests
Frontend:
- Server-rendered HTML via Jinja2
- CSS without a frontend framework
- No React, Next.js, Vite, npm/pnpm/yarn in the current checkout
Database/storage:
- SQLite file: data/gerchik_rai.db by default
- Tables: search topics, keywords, discovered groups, leads, conversations, messages, outreach events, Telegram jobs, Business threads/updates/reply jobs, Zoho sync events, logs, app settings, knowledge items, media items
- File storage next to DB:
- knowledge/ for the knowledge base
- media/ for sent materials
- telegram_sessions/ for Telethon sessions
Integrations:
- Telegram Bot API:
- Business bot webhook
- Finder bot webhook
- sendMessage, sendVoice, sendDocument, getFile, setWebhook
- Telegram user account via Telethon:
- search public groups
- join channels/megagroups
- scan group messages
- send direct messages
- mute/archive joined groups
- OpenAI:
- Responses API for AI responses
- Whisper whisper-1 for voice transcription
- TTS gpt-4o-mini-tts for voice reply
- default model: gpt-5-mini
- Zoho CRM:
- OAuth refresh token flow
- create Leads via Zoho CRM API
- Google Docs/Sheets/Slides:
- import via export URLs in txt/csv
-
8000 USD GAME AI
Gaming Apps
AI Arena RAI is a gaming platform where users create their own AI agent, upgrade it, equip it, train it, and send it into cyber battles against other agents or system opponents.
The gameplay cycle revolves around the agent: the player customizes the profile, combat style, skins, equipment, modules, and code patches that affect characteristics such as attack, defense, health, memory, reasoning, and speed. After that, the agent participates in training, PvP matches, live battles, tournaments, and club 5v5 skirmishes.
… A key feature of the game is that the battles are governed by AI logic. The agent does not simply receive a random outcome: the LLM suggests a structured action, and an internal referee/engine checks it against the rules, safety, and applies the result in the simulation. Therefore, the battle appears as a competition of strategies: the agent analyzes the arena zones, chooses actions, defends, attacks, seeks vulnerabilities, and adapts through rounds.
The game has an economy: gold, diamonds, an item shop, inventory upgrades, skins, modules, trophies, and rewards for victories. A separate layer includes live battles and betting, where players can watch matches, make quick bets, and receive calculations based on the results. Tournaments provide a higher level of competition, including elite tournaments for top agents.
For the case, this can be described as AI gaming / AI agent battle arena: the product combines gaming progression, 3D visualization, economy, tournaments, betting, and LLM-powered battle runtime into one competitive web product.
Built with Next.js, React, TypeScript, Tailwind CSS, Three.js, FastAPI, Python, SQLModel, PostgreSQL, Docker Compose, JWT authentication, OpenAI Responses API, and Polygon wallet/NFT integration.
-
80 000 USD RAI Builder
AI & Machine Learning
The RAI-Builder project is a SaaS platform for creating and managing AI agents/bots for businesses: the user connects channels, uploads a knowledge base, configures agent behavior, views dialogues/analytics/leads, manages subscriptions and integrations.
The stack is as follows:
… Frontend
React 18 + TypeScript + Vite
Tailwind CSS, Radix UI/shadcn-style components
React Router, TanStack Query, Axios
i18next for multilingual support
Recharts for analytics
Stripe.js for payment UI
Pages: landing, auth, dashboard, bot builder, messages, knowledge, analytics, audience, settings, admin panel
Python 3.12
FastAPI
Tortoise ORM + asyncpg
Aerich migrations
Pydantic Settings
JWT auth, refresh tokens, bcrypt/passlib
SlowAPI rate limiting
Babel/i18n middleware
Celery for background tasks
Data / Storage
PostgreSQL 15 as the main database
Redis 8 as broker/result backend for Celery and runtime cache
Qdrant as vector database for knowledge base / embeddings
Docker volumes for Postgres, Redis, Qdrant, and knowledge_media
AI / RAG
OpenAI
Anthropic
LangChain
Qdrant embeddings/vector search
Parsing knowledge base from web/pdf/docx/xlsx/pptx and other sources
A separate crawler service for web scraping / URL parsing
Integrations
Telegram bots via pyrofork
Telegram auth bot
Instagram Business API / Meta webhooks
Facebook Messenger / Meta Page API
Stripe billing/subscriptions/trials
Google Drive / Google Calendar OAuth integrations
SMTP / Resend / SendGrid email delivery
Web Push notifications
CRM routing / lead sync
Infra / Runtime
Docker Compose
Separate services: frontend, backend, crawler, db, qdrant, redis, celery-worker, celery-kb-worker, celery-beat, watchtower
GHCR images for prod-like deployment
Watchtower auto-update by labels
Ports bound locally to 127.0.0.1 in compose config
In short: this is not just a "bot", but a full-fledged SaaS system: React/Vite frontend + FastAPI backend + PostgreSQL + Redis/Celery + Qdrant RAG + OpenAI/Anthropic + Telegram/Instagram/Messenger/Stripe/Google integrations + Docker deployment.
-
1000 USD AI WhatsApp bot
AI & Machine LearningAI WhatsApp bot for Flipping Fund, a real estate agency in the UAE. It engages in dialogue with clients about real estate, qualifies leads, remembers context, understands requests for buying/renting/investing, and helps guide the client to the next step: selecting a property, providing materials, scheduling a viewing, or contacting an agent.
A separate strong part of the project is the processing of Property Finder leads. The bot can recognize Property Finder links, extract data about the property, retain the context of the listing, and use it in further conversations to respond about specific properties rather than generic messages.
… What is included in the project:
WhatsApp-oriented backend on FastAPI.
AI responses for real estate sales dialogues.
Memory for each chat: history, summary, facts.
Determining client intent: buy, rent, invest, mortgage, viewing, documents, handover to agent.
Property Finder parsing:
- extracting links;
- extracting lead number from WhatsApp/propertyfinder link;
- extracting property name, reference number, price, bedrooms, property type, payment plan, commission;
- saving property context in memory.
Google Drive search for materials/presentations related to projects.
Playwright automation for WhatsApp Web / Property Finder fallback scenarios.
SQLite storage for history and facts.
-
1200 USD AI HR bot TG
AI & Machine LearningAI HR bot for ISPIK, which qualifies candidates for construction vacancies in the field of solar power plants, conducts dialogue via Telegram and SendPulse Live Chat, collects questionnaire data, answers questions about vacancies/conditions, and creates a lead in Kommo CRM.
Case
ISPIK receives incoming inquiries from candidates for vacancies: installer, electrician, electrical fitter, construction equipment operator, and a female team of electrical fitters. The bot acts as the HR manager "Dmitry": it determines the language of communication, clarifies the vacancy, city, citizenship, documents, experience, qualifications, readiness for business trips, driver's license, start date, and phone number. After fully collecting the data, the bot creates a contact and a deal in Kommo CRM, and synchronizes the dialogue history in the deal notes.
-
20 000 USD Avatar TikTok Factory
AI & Machine LearningAvatar TikTok Factory — MVP platform for automated production of vertical TikTok videos with AI avatars. The system allows creating a catalog of characters, linking each avatar to Creatify, voice, content themes, and the target TikTok profile via Metricool, after which it launches the video generation pipeline: script, scenes with lip-sync, subtitles, assembly of the final MP4, and auto-publishing.
Business meaning of the project: to reduce manual production of TikTok content to a managed conveyor, where the operator configures characters, markets, themes, and schedules through the admin UI, while the backend automatically generates and publishes content in batches.
-
500 USD Bot
AI & Machine LearningVali is a production-ready WhatsApp AI assistant for a company that helps clients purchase cars/motorcycles from auctions and platforms in the USA, Canada, and Korea, and then organizes delivery and import.
-
5500 USD TikTok AI Clip Studio
AI & Machine LearningTikTok AI Clip Studio — a production-oriented MVP for automatically transforming long videos into ready vertical TikTok clips. The user uploads the source video, the system extracts audio, creates a time-coded transcription, identifies potentially viral segments, generates 3-10 short vertical videos, overlays subtitles, allows editing of titles/descriptions, and sends the video to TikTok as a draft or direct publication via the official Content Posting API.
Business value of the case: the project automates the manual work of a content editor and short-form video editor. Instead of watching a long video, manually searching for strong moments, cutting, cropping, subtitling, and preparing for publication, the user receives a managed AI pipeline from the source to TikTok-ready MP4.
… Product capabilities:
- Email/password authorization.
- Project creation for videos.
- Uploading long videos.
- Automatic transcription via OpenAI or fallback to local faster-whisper.
- AI/heuristic scoring of highlights based on factors: hook strength, emotion, clarity, novelty, sentence completeness, subtitle readability.
- Generation of vertical clips using ffmpeg.
- Subtitle overlay.
- Title/caption editing.
- MP4 export.
- TikTok OAuth integration.
- Publication in TikTok draft/direct mode.
- Monitoring pipeline jobs and logs through debug UI.
-
1500 USD Discord Call Summary Bot for A01K.
AI & Machine Learning
Discord Call Summary Bot for A01K.
Case
The service automates the processing of Discord conference calls for the trading/community product A01K. The bot connects to selected voice/conference channels on Discord, records participants' speech as separate audio fragments, transcribes them via OpenAI, compiles a short summary in Russian, and sends the finished message to the A01K internal broadcast API. The A01K system then initiates a Telegram distribution through the existing queue/pre-moderation.
…
Additionally, the service can track Discord Scheduled Events and send reminder distributions to A01K in advance, taking into account the moderation buffer.
Business Value
The project addresses the manual process of "after the call, you need to quickly write the results and send them to the traders." Instead:
The Discord call is recorded automatically;
useful content is transformed into a structured summary;
coins/tokens are highlighted separately;
the result goes into the A01K broadcast queue;
session artifacts are preserved: clips, transcript, summary;
empty/technical calls are not distributed.
-
3400 USD A01K Bot Sales
AI & Machine Learning
Project: A01K Bot Sales — Telegram Business sales assistant for A01K, which conducts incoming dialogues with current/former users and automates the renewal flow before the subscription ends.
As a case: the system addresses the task of carefully re-engaging A01K clients without manual work from the manager. The bot retrieves subscription data from A01K CRM, messages the user at D-3 / D-2 / D-1 before access ends, checks before sending whether the user has already renewed the subscription, engages in dialogue regarding objections, distinguishes between Intraday Community and Trading Campus products, issues a promo code only in the appropriate branch after a refusal at D-1, and processes payments only through the official Telegram payment bot.
Activity
| Latest proposals 10 | Budget | Added | Deadlines | Proposal | |
|---|---|---|---|---|---|
|
AI Agent
800 USD
|
|||||
|
AI agent for research
2500 USD
|
|||||
|
Marketplace
2912 USD
|
|||||
|
Game Developer
3500 USD
|
|||||
|
Looking for a Qt/C++ Desktop App Developer
2000 USD
|
|||||
|
I need a programmer for the Google Ads API.
350 USD
|
|||||
|
Development of a Classifieds and Service Showcase App with Flutterflow and Firebase
4000 USD
|
|||||
|
Website of a multi-apartment building
2751 USD
|
|||||
|
AI lead generation system
1165 USD
|
|||||
|
AI bot for news monitoring and sending to Telegram
550 USD
|
