Alexander Chadaev AI, automation and business systems
RU Discuss a project

Custom tools to manage and measure advertising

For several years, I built internal tools for a large international advertising agency: antifraud, KPI-based optimization of hundreds of campaigns across multiple platforms, install attribution, a landing page builder and advertiser discovery. I apply this experience to work for agencies, ad networks, advertisers, publishers and product teams using legitimate traffic sources.

Let's discuss your project Explore case studies

Start with a free written discussion. I respond personally within one business day. Calls are available in Russian only.

Tools I have built for advertising teams

In these projects, I usually built the entire system myself, including the interface and backend, then debugged and launched it. The case studies identify exceptions where colleagues handled mobile apps and design.

Filtering poor-quality traffic

I built custom click scoring for an agency whose traffic was poorly served by an off-the-shelf antifraud tool. I collected click and behavior data across tens of millions of clicks, developed source-specific rules and chose a filtering threshold based on business costs. After launch, problematic traffic fell by more than a factor of 20. How the filtering works.

KPI-based optimization of hundreds of campaigns

I built an optimizer for multiple advertising platforms. Rules and templates tracked KPIs and disabled ineffective segments by placement, geography, device and other criteria, much faster and more frequently than manual review. It combined data from ad networks, a tracker and the agency's analytics systems. How the optimizer works.

Reliable postbacks for iOS installs

I built a custom backend for SKAdNetwork attribution of iOS installs following clicks from websites. No ready-made solution was available at the time, and major networks supported only in-app advertising. I registered the agency's ad network with Apple and connected postback delivery to the advertiser's AppsFlyer account. This preserved the partnership and payment for iOS installs after the partner changed its terms. How web attribution was implemented.

Editing 60+ landing pages without code

I built a landing page editor for media buyers whose complex logic and analytics scripts were not supported by the off-the-shelf tools considered. I implemented access controls, logging and rollback for errors, then migrated more than 60 live landing pages. The team could edit them without changing code. How the builder works.

AI-assisted advertiser discovery

I built a system combining industry analytics, custom algorithms, web search and large language models. It regularly finds leads in selected advertiser categories, with contacts and context for outreach, helping sales identify potential partners with attractive terms. Before using it, the sales team had tried searching directly with ChatGPT and other language models, with poor results. How partner discovery works.

Competitor ad monitoring

I built an ad monitoring service with web ad scrapers, a backend receiving data from a mobile device farm, and social media collection through APIs and page scraping. I implemented deduplication, classification and a customer-facing web service with pricing plans, trial access and multiple payment methods. How the monitoring service was built.

What your team receives

  • A service integrated with ad network, tracker and BI or analytics APIs, with handling for rate limits and data-source failures.
  • An interface for media buyers or analysts to configure rules, apply templates and review action logs.
  • Data-source availability monitoring and alerts when connections fail.
  • System documentation and source code in your repository.
  • Training and help migrating existing campaigns to the system.

AdTech projects delivered

A large advertising agency with international operationsIn production

Custom fraud detection cuts problem traffic more than 20-fold

The agency bought traffic from multiple sources but still received advertiser complaints despite an off-the-shelf filter. I analyzed click data and tested hypotheses to find what it was missing, then built a custom click-scoring system.

>20xreduction in problem traffic after launch; major advertisers retained Read case study
Large international advertising agencyIn use, with ongoing development

Automated ROI optimization across hundreds of campaigns and 10 ad networks

Manual checks could not keep up with declining placement quality. I built a system that combines network, tracker and agency analytics data to disable underperforming segments according to media buyers' rules.

Hours saved weeklyby automatically checking hundreds of campaigns and applying media buyers' rules Read case study
A large international advertising agencyIntegration launched, payment requirement met

Keeping iOS installs payable with SKAdNetwork for web ads

A key partner would pay only for AppsFlyer-confirmed installs. No ready-made SKAdNetwork solution existed for website ads at the time, so I built a custom integration to provide the required confirmation.

Install payments preservedThe agency retained its key partnership and payment for iOS installs under the new terms Read case study

All case studies

From reviewing data to daily use

  1. 1

    Understand your workflow

    I review the ad networks and trackers you use, how you build reports, and where your current processes waste time or money.

  2. 2

    Collect data for decisions

    I connect data sources and set up statistics collection. For traffic-quality work, I investigate observed clicks so filtering rules are based on evidence rather than assumptions.

  3. 3

    Test the working version

    We introduce the tool on a subset of campaigns or traffic, compare results with the manual process and make the necessary fixes.

  4. 4

    Expand use of the system

    I help migrate the remaining campaigns and train your team. The system can then serve as a foundation for automating further tasks.

Ways to work together

Cost depends on scope, integrations, data quality, load and system requirements. After discovery, I estimate the next stage. If development of the first version starts within one month after discovery is completed, I credit the discovery fee toward it.

  • Discovery and planning

    I review workflows, available data, and ad network, tracker and BI or analytics APIs. You receive a preliminary plan and an estimate for the next stage, whether development or a traffic-quality assessment. Where useful, I include a high-level diagram or illustrative prototype, not detailed specifications or a finished prototype.

    Budget: From $2,500

    Timeline: 5-⁠10 business days

  • First working version

    I build the agreed AdTech solution, from analytics and integrations to campaign automation. We define the first version's scope in advance; I quote and deliver larger projects in separate phases.

    Budget: From $15,000

    Timeline: 4-⁠8 weeks

  • Ongoing support and development

    I add features, rules and integrations, connect new ad networks and adapt the system to API changes. We agree on scope, availability and incident response in the contract. Major additions are quoted separately.

    Budget: From $4,000 per month

    Timeline: Monthly

A paid-traffic quality assessment can be a standalone project: collecting statistics, identifying poor-quality sources and analyzing losses. In the agency project, problematic traffic fell by more than a factor of 20 after my antifraud system launched. That is a result from one project, not a promise for every traffic source.

Teams and projects I can help

A good fit

  • Agencies, ad networks and publishers using legitimate traffic sources.
  • Advertisers, product teams and mobile app teams needing paid-traffic quality analysis and end-to-end analytics.
  • Marketing and sales teams whose workflows need custom automation tools.

Projects I do not accept

  • Promotion of gambling, betting or questionable financial offers.
  • Tools for cloaking, evading moderation or circumventing platform rules.
  • Solutions designed to mislead users.

What to know before we start

Can you integrate with our tracker and BI tools?

Yes, I will review the available integration options. The optimizer I built combined data from ad networks, a tracker and the agency's analytics systems. Unreliable or missing APIs, rate limits and platform delays do not necessarily rule out integration, but they need to be investigated and accounted for in the design.

How would you build antifraud for our traffic sources?

I start by collecting click and behavior data to identify the problematic traffic coming from your sources, then develop rules based on those findings. We choose the filtering threshold by weighing the cost of letting poor-quality clicks through against the loss of real customers. Stricter filtering is not always more profitable.

Can we work under an NDA?

Yes, I sign confidentiality agreements. These obligations also apply to the published AdTech projects, so the case studies do not name the companies or show screenshots of internal systems.

What is your experience with iOS install attribution?

I registered the agency's ad network with Apple for SKAdNetwork and implemented the postback chain to AppsFlyer for installs following clicks from websites. Different versions of the documentation did not explain the failures. Each system's support team covered only its own part, leaving no one to troubleshoot the full chain. Apple did not help in time, so I debugged the integration independently. I have also developed server-side postbacks for mobile apps.

How do we communicate during the project?

I work remotely. We can start with a free written discussion in English, and I respond within one business day. Project correspondence, documentation and progress updates can be in English; live calls are in Russian. We agree on the contract and payment terms before work begins.

What would you like to solve?

Share a detailed brief, requirements or any materials you already have. If you have none yet, a few sentences about the current situation and your goal are enough. I review every message myself and respond within one business day.

  • You will work directly with me, not a sales representative.
  • A free initial written discussion will help establish whether I can help and where to start.
  • If the project is outside my expertise, I will tell you directly.
Alexander Chadaev

Alexander Chadaev

Telegram @chadaev_ai hello@chadaev.ai

Include timing and supporting files

Attach requirements, sample documents, screenshots or other supporting materials. Maximum 10 files, up to 20 MB in total.

Your form details are processed to handle your enquiry as described in the privacy policy.

Let's discuss your project Telegram