From Real Projects to Articles, Videos, and Social Posts by Fedor DorodnovFrom Real Projects to Articles, Videos, and Social Posts by Fedor Dorodnov

From Real Projects to Articles, Videos, and Social Posts

Fedor Dorodnov

Fedor Dorodnov

Useful project experience often stays in code, documents, and task boards. Sharing it means finding an angle, checking the details, writing, and making visuals. This AI-agent system brings those steps into a repeatable workflow, helping engineers and teams turn their experience into content people can learn from.
I designed and built the system's agent roles, instructions, scripts, and workflow, including API integrations and prompts for image generation. The work also involves refining how agents review each other's output and when a draft needs another pass.
The process starts with a topic and an audience. Agents examine the relevant code and project documents, prepare a draft, and check its technical claims against the sources. English and Russian each get a separate review for natural language and clarity. Unsupported claims and awkward wording go back for revision, with the final editorial decisions kept in human hands.
Project evidence, drafting, technical and language review, and revision before approval.
Project evidence, drafting, technical and language review, and revision before approval.
A project can then become an article, a short social post, a portfolio case study, or an animated video, with copy and visuals suited to each format. Technical terms get their own explanations where needed. Tasks, source material, and editorial decisions stay together, so the next piece can build on work already done.
Published output already includes articles, posts with animated videos, and a School Platform case study on Contra.
On LinkedIn, the results include:
6,717 impressions, 62 reactions, 20 comments, and 3 reposts on a Russian post about job searching.
299 impressions, 3 reactions, and 1 comment on its English version.
11 to 40 impressions on six earlier technical posts and article announcements.
Selected LinkedIn post results: impressions, reactions, comments, and reposts.
Selected LinkedIn post results: impressions, reactions, comments, and reposts.
For engineers and teams, the value is a practical way to share expertise regularly, even when writing is not their strongest skill. It supports a personal brand built around work they can explain and demonstrate. Trying different topics and formats, then reviewing the response, helps shape what to publish next.
I'm now bringing that approach to Contra, exploring which project stories connect with people here and how to present them through portfolio Works.
Like this project

Posted Oct 1, 2026

An AI-agent system for turning project experience into useful content, building a personal brand, and testing topics and formats.