Andrii · AI Automation Specialist
AI Automation Specialist Ukraine · remote · full-time or project-based

I help businesses work
faster and spend less.

I take manual routine off the team with automation and AI: I connect systems to each other and build custom tools instead of pricey subscriptions. The result shows up in numbers: man-hours saved, service payments cancelled, processes that run on their own. Delivered end to end: AI agents and bots, parsers and monitoring, high-volume operations, internal dashboards, API integrations. My specialisation is not an industry, it is a type of task: 2 years, 10+ different domains — from marketing agencies to crypto analytics.

AI agents, bots and RAG Integrations and data consolidation Parsers, monitoring and alerts High-volume operations, multi-account Internal tools and dashboards
Flagship caseFor a SERM agency I built lead generation and internal automation from scratch: website, parsing, outreach, bots, task management, team workflows. Within a few months — 25+ clients, a team that spends less time and gets more done, plus $3,600 a year saved on software they would otherwise have to pay for.
25+ clients
brought to an agency by the lead-gen pipeline I built from scratch
11M+
contacts in a platform with semantic search, filters and AI classification
2,000+ / mo
high-volume browser operations on a conveyor, with no human in the loop
2,500 accounts
a fleet I ran with the full infrastructure behind it — 1.5 years
Cases with numbers Message me on Telegram Two cases come with a “Proof” button — it opens a chat with the actual client.
01 Profile

I build through an AI pipeline (Claude Code, Codex, MCP): the AI writes the code — I own the problem statement, the architecture, the tests and the acceptance. Solutions run on LLM APIs (Claude, OpenAI, Gemini) + Python + REST APIs and webhooks; prompt work is daily practice — design, prompt chaining, testing model output. A typical internal tool reaches production in days, not sprints.

My specialisation is not an industry but a type of task — that's why I get into a new domain fast: I start by mapping the team's process together with them, and pick the solution to fit it. The clearest example is a SERM agency: they came with no spec, just “look at where it hurts”. We mapped the processes together and closed them stage by stage — lead gen, outreach, the website, a task manager, dashboards, bots — until marketing, sales and operations ran on their own. Over 2 years I've entered 10+ domains this way: marketing and reputation agencies, media buying, crypto analytics, paid communities, sales.

My strengths are autonomy and predictability. I take the task from a conversation and return a written specification, then work in phases: progress is visible from outside and you can stop at any point. My main professional habit, going back to running a 2,500-account fleet, is to verify the outcome after every action instead of trusting that the script worked. And I count the benefit in numbers: person-hours removed, subscriptions cancelled. “Done” for me means “running in production”, not “the code is written”.

The same “trigger → processing → action” workflows I build in custom code — no platform limits and no subscription per step; at volume that is a difference in money. If part of your processes already lives in n8n / Make / Zapier, I'll pick it up, get up to speed in a day or two and tell you what is worth moving into code and what is cheaper to leave as it is.

02 Key cases

What has already been built for teams

The numbers below are person-hours and money taken off the team — not the client's revenue. For two cases you don't have to take my word: the “Proof” button opens a Telegram chat with the client, ask them yourself.

Business processes · flagship Proof — ask the client

An agency automated end to end: marketing, sales, operations

The team came with no spec — just “look at where it hurts”. I went through their processes and closed the whole loop, from acquiring clients to the team's daily work. This is the fullest example of how I work: not one automation, but a system of connected parts.
  • Lead generation from scratch: automated collection of target audiences in Telegram + warmed accounts + outreach → 20+ clients; company collection across 15+ niches from open sources + email outreach → ~5 more clients
  • Custom outreach and parsing software instead of a paid analogue at $300/mo — that is $3,600 a year and no platform limits; I later reused the same software in other projects
  • Landing page with a service calculator: the request lands in the admin's Telegram instantly, and the calculator removes the “let me price it” step from the manager and filters out non-buyers before the conversation
  • Internal task tracker, 24/7: tasks moved out of chat threads into one place — project, deadline, assignee, “who delivered and who didn't” analytics. Instead of “it got lost somewhere in the chat” — a visible status
  • Control dashboard: conversion of every campaign and channel is visible — decisions on numbers, not on gut feel
  • Telegram business bot on the official Business connection: control of working accounts and process notifications for the team
  • A website with SEO pages that brought requests in the first weeks
25+ clients$3,600/year saved15+ niches parsedfull loop, end to end
Result: 25+ clients brought in by a lead-gen pipeline I built from scratch, plus $3,600 a year saved on a subscription. Marketing, sales and daily operations of one agency — automated, with one person owning the whole contour.
AI agents · RAG

First-line AI assistant on the company knowledge base

A closed paid community: people kept asking the same questions, and human support simply could not cover 24/7.
  • Answers from internal sources — database, website, articles — plus web search when the base has no answer; remembers the whole conversation
  • Multimodal: listens to voice messages and replies by voice, recognises images, analyses documents, prepares summaries
  • When it doesn't know, it doesn't invent — it hands over to a human
24/7instant answersfirst line without a human
Result: people get an answer instantly, at three in the morning too. The “explain the same thing for the tenth time” routine is off the team; a human steps in only where the bot doesn't know.
AI agents · sales, CRM

AI sales assistant that keeps the CRM up to date

First contact with clients ate the managers' time: half of the conversations never reached a deal, but each one still had to be worked and logged in the CRM.
  • Writes to clients, advises from the script and the knowledge base, qualifies interest and creates the CRM record itself
  • Trained on six months of the team's real chats — it answers the way the team answers, not “like a chatbot”
  • Escalates to a human manager exactly when a human sale is needed
trained on 6 months of chatsqualification on the botCRM fills itself
Result: repetitive questions and qualification sit on the bot, the manager joins an already warm lead. Manual CRM data entry after every conversation disappeared as a separate job.
Monitoring · alerts

Event monitoring with instant notifications

The team had to react to news about tracked projects within hours and be first — too many sources to check by hand.
  • Parsing per object from several sources at once: Twitter, Telegram, Discord (specific threads), websites — fires almost instantly
  • An event appears → the subscriber gets a Telegram notification right away; alert thresholds are configurable
hours → minutes3+ sources in one feed
Result: going through sources by hand used to eat hours every day — it became one feed with alerts and a reaction time in minutes. It only pings when there is actually something to react to.
Data · infrastructure

Data storage and processing platform — 11M+ contacts

The team had accumulated scattered client databases — sitting in files across folders and in chat threads, i.e. dead weight.
  • Deployed a server and a platform holding 11,000,000+ contacts (emails, phones, names)
  • Import, export, filtering and search; under the hood AI recognises and classifies contacts, filtering out duplicates and junk
  • Integrated a vector database: contacts and bases are searched by meaning, not just by exact field match — the platform finds thematically close bases on its own
  • Built in an AI assistant: it suggests which base fits the task, helps assemble the right slice and hands it over ready for export
11M+ contactsa slice in secondssemantic searchAI classification
Result: the slice you need is found and exported in seconds instead of digging through files by hand. Dead databases became a working tool for the team.
High-volume operations

A browser operation conveyor — 2,000+ per month

A large corporation needed one repetitive registration operation at a volume of 2,000+ per month. Ready-made purchased accounts were cheap but died within days — the process kept stalling.
  • Full cycle automation (ZennoPoster): correct user emulation, handling captchas and anti-bot checks, reliable proxies and quality mailboxes as consumables
  • Daily batches on a schedule, with the outcome of every operation logged
2,000+ operations/moresults survive for weekszero manual work
Result: the need for 2,000+ quality operations per month is covered consistently, with no human involved. The same conveyor transfers to any resource you burn at volume: ad accounts, domains, mailboxes, payment profiles.
Browser automation · with outcome verification Proof — ask the client

Publishing content at volume instead of a team of operators

A reputation marketing agency needed to publish reviews and file complaints at volume. Normally a group of people does this by hand in anti-detect browsers with proxies — slow, expensive and dependent on whether the operator showed up for the shift.
  • Full cycle with no human in it: browser fingerprint, proxies, captchas, publication with the given text and rating
  • Behaviour randomisation: different entry points, different route through the site, different pauses and actions — the session doesn't look templated
  • Survival check after 24 hours: the script comes back on its own and verifies the result is still live. That is quality control, not “fire and forget” — you see the real output, not the number of attempts
  • Daily report to a Telegram bot: how many went through in 24 hours, with no manual spreadsheets
up to 50 publications/day~8 person-hours/day≈ one operator salaryoutcome verified after 24h
Result: up to 50 publications a day — that is ~8 person-hours of manual work removed entirely, a full operator position; plus ~30 complaints a day (another 1.5–2.5 hours). The ceiling was held by the platform's risk limit, not by the software: throughput scales with the number of threads.
Web3 / crypto · 1.5 years — the foundation I came into development from

Browser automation and Web3 analytics: 30+ projects, a fleet of 2,500 accounts

A year and a half in crypto: I was responsible for a fleet of 2,500 accounts, automated every browser action instead of a team of operators, and ran analytical market monitoring. This is where my sense of scale, the anti-detect infrastructure skills and my main professional habit come from — verify the outcome after every action instead of trusting that the script worked. From that foundation I moved into AI automation of business processes.
  • Automation for 30+ projects (ZennoPoster): the full cycle from registration to daily routines — interaction with DEX platforms (swaps, cross-chain bridges), prediction markets, trading scenarios, testnet activity
  • A fleet of 2,500 accounts: each with its own wallet, browser fingerprint and proxy. Behaviour randomisation — different routes, pauses, amounts and order of actions, so sessions are never templated
  • The infrastructure behind it: anti-detect browsers, proxies, captchas, wallet accounting and maintenance, balance and status monitoring — so the fleet stays alive instead of “something broke and nobody noticed”
  • Analytics and research: daily monitoring of 100+ sources — blockchains, DeFi protocols, exchanges, infrastructure projects. Digging into teams, funding and mechanics, reacting to events fast
  • My first parsers were built right here: pulling funding and activity data from websites and blockchain explorers into one picture instead of manually cycling through dozens of tabs
30+ projects automatedfleet of 2,500 accounts~1,990 accounts on one project$1.3M+ in rewards on that fleet100+ sources monitored
Result: the strongest case of that period — a project where automation carried ~1,990 accounts; total rewards on that fleet exceeded $1.3M. The whole infrastructure underneath — wallets, proxies, maintenance, health control — was mine, both the setup and the daily upkeep.

Also built: Telegram account management software (a proxy per account, spam-block checks, warm-up); a bot that transcribes meetings and voice messages and extracts tasks from them; personal bots for my own routine — monitoring the status and limits of the AI services I work with; bots for handling dozens of crypto wallets — balances, statistics, analytics.

03 Where it pays off

What I can automate in your team

Whatever the industry, manual hours go into the same places: data is consolidated by hand, repetitive requests are answered by people, incoming documents are read with human eyes, and internal tools sit in a queue behind the product. Below are eight of those places and exactly what I take off them. Each one is marked with the kind of experience I have.

Reporting is assembled by hand from several systems
Every morning someone exports data from ad accounts, the CRM and spreadsheets and merges it into one report — day after day. I collect it automatically: pull from APIs and sources, merge into the breakdowns you need, deliver as a dashboard or a Telegram report, plus an alert when the numbers drift from target.
✓ direct experience: control dashboards, parsers and API integrations, merging data from different sources
Repetitive requests are handled by a person
Clients and employees ask the same things, while the answer already sits in the knowledge base, the internal policies and old chats. I put a first-line AI assistant on your own sources: it answers from the base, keeps conversation context, and hands over whatever it doesn't know — without inventing answers.
✓ direct experience: RAG assistant for a paid community, multimodal (voice, images, documents)
The CRM is filled in by hand after every conversation
A manager spends part of the day not selling, but typing in what was already said in the chat. I automate first contact and qualification: the assistant runs the conversation from a script, assesses interest, creates the CRM record itself and passes a warm client to a human.
✓ direct experience: AI sales assistant with CRM handling and escalation to a manager
Incoming documents and forms are read by eye
Invoices, contracts, questionnaires, CVs, screenshots — someone opens every file, pulls out the fields and copies them into a spreadsheet. I build a conveyor: AI reads the file, extracts the fields you need, classifies it and puts it into structure, leaving edge cases for a human to check.
✓ direct experience: AI classification of 11M+ contacts, document and image analysis by bot
Finding clients and partners is manual labour
Someone goes through directories, websites and chats by hand, collects contacts into a spreadsheet and writes to each one. I set up a pipeline: collection by your criteria from open sources → cleaning and classification → personalised sending → reply tracking in a dashboard.
✓ direct experience: 25+ clients for an agency, 15+ niches parsed, custom outreach software
Problems are discovered too late
A service went down, a link stopped leading where it should, a competitor changed prices, important news broke — and you find out after the fact. I set up monitoring: source checks plus a bot walking the critical path every N hours, with a Telegram alert the moment something breaks.
✓ direct experience: multi-source monitoring with instant alerts, full-cycle browser checks
Repetitive operations where there is no API
Some systems simply have no integration — so people click through them hundreds of times: registrations, exports, form filling, status updates. I close that with browser automation that verifies the outcome after every action and reports daily.
✓ direct experience: 2,000+ operations/mo on a conveyor, a fleet of 2,500 accounts, outcome survival checks
Internal tools sit in a queue behind the product
The team needs a tracker, a dashboard, a bot or data exchange between systems — but engineering is busy with the product and operations waits for months. I take on what is around the platform: task trackers, bots, dashboards, integrations. I don't rewrite your product.
✓ direct experience: 24/7 task tracker, business bot, dashboards, a platform with 11M+ contacts
04 How I work

A tuned process, not vibe coding

Every project goes through the same path — that is why the result is predictable and the progress is visible from outside.

Briefing. I take the task apart in detail, gather the whole context and keep asking questions until it is unambiguous.
Specification. I write down what we are building and how we will verify it is done.
Phased plan. Progress is visible and you can stop at any point.
Implementation. AI pipeline, parallel agents; complex things are assembled over days up to a working version.
Tests and review. Automated tests plus mandatory independent code review (Codex among others) — so it is architecturally right, not just “it runs”.
Launch and handover. Deployment (VPS, Docker, systemd) plus clear instructions for the team. “Done” means it works in production.
05 Skills and tools
Core tools
Claude CodeCodexAI-assisted developmentprompt engineering (prompt chaining, optimisation)Python (paired with AI)ZennoPoster
AI and bots
LLM APIs (Claude, OpenAI, Gemini)AI agents and MCPRAG / knowledge basesvector databases / semantic searchbuilding skills and plugins for agents (SKILL.md, subagents)Telegram bots (business bots, user bots, assistants)STT/TTS (Deepgram, ElevenLabs)image and document recognition
Data and analytics
parsers and scrapersdata collection, cleaning and classification (ETL)sources: websites, Twitter, Discord, Telegram, blockchain explorersXHR interceptiondatabases (SQL, vector)data and config formats (JSON, YAML)control dashboards
Browser automation
user action emulationanti-detect browsers (Dolphin, AdsPower)proxiescaptchas and anti-bot checksmulti-account (2,500 accounts)
Integrations and infrastructure
REST APIs / webhooksintegrations with CRMs and internal systemsGoogle services (Sheets, Docs)OAuth / service accountsgit / CILinux VPSDockersystemdCloudflareserver deployment
06 Experience
AI Automation Specialist — freelance and projects for teams
early 2026 — present

I automate processes for agencies, businesses and teams: AI agents and bots on LLM APIs, parsers and monitoring, integrations via REST APIs and webhooks with CRMs and databases, internal services and dashboards. Full cycle: process analysis, specification, plan, implementation, tests, production deployment (cases above).

Browser automation, multi-account and analytics — Web3, in a team
summer 2024 — early 2026

The year and a half I came into development from: automation for 30+ projects, running a fleet of 2,500 accounts with the full anti-detect infrastructure, high-volume registrations, and analytical monitoring of 100+ sources with my first own parsers.

Education

Civil engineer — Prydniprovska State Academy of Civil Engineering and Architecture
Courses: AI-assisted development, web design (Figma)

Languages

Ukrainian · native
Russian · fluent
English · Intermediate (fluent reading, written communication)