Kristina Chernodub LinkedIn

I’m a Product Manager

I work hands-on with tools like Codex, Claude Code, and n8n in my day-to-day workflow.

I do my best work in under-structured environments where product judgment matters more than process for its own sake.

  • 6 years in digital and IT
  • 28 projects delivered
  • 14 clients
  • Hands-on AI practice
LinkedIn Telegram
Kristina Chernodub portrait

Work

Selected work

AI Meeting Workflow Agent

Post-meeting PM work dropped from 60+ minutes to ~30, saving ~18 hours per month across parallel projects

  • OwnedI owned the product layer: identified the bottleneck, wrote the specification, designed the context layer and selected 2 automation scenarios, authored the project knowledge base, and set the evaluation criteria before implementation started.
  • Why it matteredThe time savings are only real if the output can be trusted. An agent that drafts wrong tasks or misreads decisions costs more time than it saves. The rubric and HITL design were what made the numbers count.

n8n, OpenAI API — connected to meeting transcription, project management, and knowledge tooling

Read the case study →

Delivered 28 projects for 14 clients, 3–5 in parallel. 8 came back for follow-on work
  • OwnedI owned discovery through release: clarifying what clients actually needed before scope was locked, turning that into user scenarios and acceptance criteria, and keeping technical and release decisions visible across parallel teams.
  • Why it matteredClients usually knew what they wanted to build, not what the product needed to do for users. Getting that clear before implementation started was what kept parallel work manageable and recoverable.

Jira, Notion, Figma, GA4, GTM, Looker Studio

Reached 500+ monthly SEO visitors and 1.75% conversion before shutting it down cleanly
  • OwnedI researched the niche, built the offer, automated delivery, and tracked the numbers end to end.
  • Why it matteredThe useful part was seeing that the economics mostly never worked and making a clean decision instead of dragging the project forward.

Tilda, Make.com, Google Sheets, Notion API, email automation, Telegram bot

About

A few things that matter in how I work

I like work that becomes clearer, cleaner, and more useful over time.

Start from the real bottleneck

I want to know what is actually wasting time, breaking trust, or making the work harder before I propose a process or a tool.

Build context early

I write things down early and keep context reusable because good documentation is memory, alignment, and faster decisions later.

Stay close to the outcome

I care about delivery, but I also care about whether the work still makes sense for users and for the business.

Keep ownership visible

I prefer transparent logic, visible tradeoffs, and work that does not turn into a black box.

AI

I use AI where it earns its place

I use AI to structure context, reduce repetitive work, and test workflows faster. What matters to me is not the demo, but whether the system stays useful under real constraints.

  • Context matters more than prompting

    Useful output usually comes from better context, clearer boundaries, and cleaner inputs.

  • Human review stays in the loop

    I use agents to draft, test, and reduce follow-up work. I do not outsource judgment.

  • Real use is messy

    If nobody mentions friction, verification, or failure modes, I trust the story less.

Activity

Live Apple Health snapshot

Workouts, sleep, steps, and consistency from Apple Health. Reading Ray Dalio's Principles was part of what made me care enough about transparency to keep this public.

Contact

If this feels relevant, reach out

I'm open to Product Manager and product-adjacent roles, and to consulting conversations around practical AI workflows. If you're already building with AI or want to start but need a clearer way in, I'd be glad to talk.

Connect on LinkedIn Message on Telegram