Skip to content
PD
Pavel Duglas - AI Automation & MVP Architect
153 client reviews from freelance platforms

Automation example demo

WebsiteParserSheetTelegram

Rows collected: 1,249 Alerts sent: 38

MVP example demo

IdeaPrototypeLaunchUsers

Sign-ups: 128 Paid: 9

9

years of practice, since 2017

153

client reviews

30

case studies

30+

systems launched

164

wiki articles, 11 guides

"Very fast parsing, thank you! It even returned a few more numbers than expected, I recommend him to everyone. I have ordered twice now, happy with all of it, and I will be back."

sotasoftdv · Kwork · Translated from Russian

"Thanks to Pavel, the task is done. Always reachable, gave me detailed instructions and a guide, I will come back and I recommend him to everyone."

MarkBorisov · Kwork · Translated from Russian

"Pavel Duglas is an excellent specialist, he showed me how to build a trading bot in a few clicks right inside BAS. Recommended, it is worth it."

IvanSorokin787 · Kwork · Translated from Russian

Featured Work

Selected projects and case studies

Real systems: AI bots, SaaS platforms, CRMs, finance dashboards and automation.

All Projects
The finance overview: net worth, cash and banks, crypto, monthly income, spending, net flow and savings rate, cash flow by month, spending by category and the largest operations
Web Apps

Duglas Finance OS - a Personal Capital Management System

All the money in one place: cash, banks, 17 crypto exchanges and cold wallets in one base currency, goals with a savings pace, a journal of yield positions and hints on where idle capital could work.

Coverage
17 exchanges
Volume
8,389 operations
APIs of 17 crypto exchanges and T-BankDefiLlama and exchange earn dataDeFi position scan across walletsExchange rates from open.er-api.com +2
A generation result in the VK chat: for the prompt «round island» the bot sent a photorealistic island in the ocean, plus a hint on switching models and finding the balance
AI Products

Neurox - AI Photo and Video Generation Bot for VK

A bot for a VK community that turns text and photos into images and video: 12 AI models in one chat, payments through YooKassa and an owner's panel with the revenue, cost and margin of every generation.

Modes
4 modes
Coverage
12 AI models
VK APIAI model aggregator (one API)YooKassa: SBP and SberPayGeneration queue on a worker +2
The LeadGen Outreach dashboard: the worker is running, today's proposals and leads, calls this week, leads that need a follow-up and the daily proposal limits across 12 platforms
Automation

LeadGen Outreach - a Client-Finding System for Freelance Marketplaces

My own CRM with an AI engine: 14 sources of work, task filtering, proposals within each platform's limits, a pipeline up to prepayment, AI reply drafts and a demand analysis across 5,811 tasks.

Coverage
14 sources
Throughput
589 proposals
AIML API: gpt-4o-mini and gpt-4otext-embedding-3-small embeddingsSemantic task clusteringTelegram Business and MAX +2
Emaze Tours editor: a tour on a light canvas with a slides panel, a toolbar (image, video, text, HTML, shape, frame, link), text-formatting controls, a placed image, text objects and a blue ellipse, and a right-hand settings panel with a 3D room-type selector
Web Apps

Emaze Tours - Interactive Virtual Tour Builder

A Canva/Emaze-style web app where anyone builds an interactive 3D virtual tour in a drag-and-drop editor, then publishes it to a shareable link or embeds it via iframe.

Next.js 16 (App Router)React 19TypeScript (strict)Three.js + React Three Fiber + Drei +6
WBOzonMonster settings: Wildberries/Ozon service selection, the profiles folder, and operating modes - buyouts, pickup point, feedback, warming, receipt collection, profile creation, cart clearing
Automation

WildberriesOzonMonsterBot - marketplace promotion automation

A multi-mode BAS bot for a marketplace promotion team on Wildberries and Ozon: profile creation and warming, auto-buyouts, pickup-point data collection and review handling - up to 200 threads instead of manual work.

Scale
200 threads
Speed-up
4-5x
Browser Automation Studio (BAS)Multithreading (200 threads)Anti-detect profilesSMS registration +2

What I Build

Systems that turn ideas into working products

Complete digital systems: interfaces, databases, automation and AI logic, deployed and running.

Client reviews

What clients wrote after the work shipped

Transcribed word for word from the review pages of the marketplaces I work through. Each card links to the screenshot it came from.

See all 153 reviews

Very fast parsing, thank you! It even returned a few more numbers than expected, I recommend him to everyone. I have ordered twice now, happy with all of it, and I will be back.

s sotasoftdv Kwork

Phone database from Drom. Parsing the Drom car listings

Translated from Russian

Pavel is a great specialist, he delivered the parser in a few hours. I will keep ordering from him. Huge thanks for the professionalism and the speed.

v volodymyrlemchenko Kwork

A BAS bot to automate routine tasks

Translated from Russian

Thanks to Pavel, the task is done. Always reachable, gave me detailed instructions and a guide, I will come back and I recommend him to everyone.

M MarkBorisov Kwork

Setting up and launching the VK -> Telegram/MAX autoposting service

View Case Study

Translated from Russian

Genuinely a master of his craft. I learned a lot of new and useful things about Telegram outreach. I had burned a lot of money paying intermediaries before this, and after talking to him I will be doing it all myself.

v viktorTR Kwork

Telegram software setup

Translated from Russian

Pavel Duglas is an excellent specialist, he showed me how to build a trading bot in a few clicks right inside BAS. Recommended, it is worth it.

I IvanSorokin787 Kwork

A TradingView crypto trading bot built on BAS

Translated from Russian

Quality matches the price. All great, done at a really fast pace. I will be back.

S Sokia Kwork

Parser and wallet checker for NFT collection holders on context.app

Translated from Russian

How I Work

From first call to working product

A predictable process with visible progress at every stage, and a scope agreed in writing before the work starts.

  1. 1

    Logic & Value

    We define the business problem, the core value and the riskiest assumption. No code yet - clarity first.

  2. 2

    Architecture

    I design the system: data, interfaces, integrations, AI logic. You see the map before paying for the territory.

  3. 3

    Build

    Short iterations with visible progress. A working version appears early; feedback shapes every week.

  4. 4

    Launch & Scale

    Deploy, measure, automate, extend. The system grows only after it proves its value with real users.

From the Blog

Practical notes on AI, automation and MVPs

All Articles
AI Automation 7 min read

Stop Scoring AI Output by Vibes. Measure How Much Humans Fix It

Most AI features are judged by demos and gut feeling. I track one thing instead: how much a human has to change the output before it ships. Here is how to capture it, compute it and use it to pick models and prompts.

Read Article
AI Automation 7 min read

PDF Is Not Text: Build an Ingestion Layer Before Your LLM Sees It

Most bad RAG answers come from broken PDF extraction, not a weak model. Here is the ingestion layer I build: triage, coordinate-based extraction, structure rebuilding, per-page OCR fallback and quality gates.

Read Article