Case Studies Enterprise Contact Download CV

Pavel Kuzko

AI Product Manager

I take AI from prototype to production in regulated industries.

148K
Users
£500K
ARR
3
AI Products Shipped
15+
Years in Fintech
View Case Studies Download CV LinkedIn
Pavel Kuzko

Skills & Tools

What I Work With

Product Management

Product Strategy Roadmapping A/B Testing User Research Agile / Scrum Jira Figma PRDs

AI & Engineering

Python SQL LLM Integration NLP ML Pipelines Prompt Engineering OpenAI API Anthropic API

Infrastructure & Data

Docker PostgreSQL Redis REST API Grafana CI/CD n8n Make.com

Case Studies

Proof of Work

Each project demonstrates a specific capability — from 0-to-1 product builds to enterprise-scale delivery.

TgSight Insight Bot — AI-powered user interest classification across 12,000+ categories
Featured Case Study

TgSight — AI Analytics Platform

Problem

Telegram has 1B+ users but zero built-in analytics. Channel owners had no way to understand audience demographics or segment users.

What I Did

  • Joined as first Product Manager, owned vision and roadmap from day one
  • Designed ML classification pipeline for automated interest detection
  • Built 3 products 0→1: Interest Profiler Bot (118K users), History Bot (28K users), Channel Insight Bot (2K users)
  • Managed distributed team of 12 across multiple time zones

Results

  • 148,000+ MAU (B2C)
  • £500K ARR
  • B2B clients in fintech and marketing
Python Django Celery PostgreSQL Docker ML Classification NLP
Vapi AI Voice Workflow — multilingual call qualification and lead routing
AI in Regulated Industry

AI Voice — Sterling Law (Legal 500)

Problem

Legal 500 immigration firm (£5M+ revenue) was losing leads through manual call screening. Multilingual calls (EN/RU/UA) were misrouted or lost.

What I Did

  • Hired as first technical product hire — sole owner of AI strategy
  • Designed multilingual voice flow: language detection → qualification → lead scoring → CRM routing
  • Integrated with Clio (legal practice management), Asterisk PBX, and internal CRM
  • Built AI-assisted case management workflows

Results

  • Eliminated manual call screening entirely
  • 60% reduction in manual processing
  • 30% faster case turnaround
  • Reduced cost-per-qualified-lead
Vapi.ai LLM Integration Clio API Asterisk PBX Python Workflow Automation
xR2 Prompt Editor — version control, variables, multi-model testing
Live Product

xR2 — Prompt Management Platform

Problem

AI teams hardcode prompts into codebases. Every change requires a deploy. No way to A/B test or measure which prompt drives revenue.

What I Built

  • Multi-tenant architecture with role-based access
  • Prompt versioning lifecycle: Draft → Staging → Production
  • Real-time A/B testing engine with statistical significance tracking
  • Event tracking with conversion funnels and ROI attribution
  • REST API with multi-LLM provider support
FastAPI PostgreSQL Redis Docker REST API n8n Make.com
Try xR2
AI CLIENT PIPELINE LEAD INTAKE Website, Referral, Ads ELIGIBILITY CHECK ML classification model AFFORDABILITY SCORE ML scoring model QUALIFIED LEAD NOT ELIGIBLE AUTO NURTURE CAMPAIGN Automated emails + follow-ups • Conversion 8% → 18% CLIENT ONBOARDED 80% automated £200K saved
AI Automation

Bug Free Immigration — AI-Powered Client Pipeline

Problem

Immigration consultancy spending 40h/week on manual onboarding. Eligibility and affordability checks done by hand. Conversion stuck at 8%.

What I Did

  • Designed AI system for full-cycle client onboarding automation
  • Built ML-based affordability and eligibility scoring models
  • Automated 80%+ of client pipeline workflows
  • Implemented automated nurture campaigns

Results

  • £200K saved in operational costs
  • Manual work: 40h/week → 8h/week
  • Conversion rate: 8% → 18% (+125%)
  • Lead qualification accuracy improved by 45%
Python ML Scoring LLM Integration Workflow Automation CRM Integration

Enterprise Background

Proven at Scale

Before building AI products, I managed digital products and delivery teams at enterprise scale.

Enterprise

Sberbank

2016 – 2018

Product Manager, Digital Banking

Scale: 280M+ users (Russia's largest bank)

  • Owned product vision, roadmap, backlog, and cross-functional delivery
  • Coordinated across engineering, design, and business teams
  • Shipped digital banking features to 280M+ user base
Enterprise

TransContainer

2022 – 2023

Delivery Lead

Russia's largest container operator

  • Led delivery team of 10
  • Migrated monolith → microservices: feature delivery 3 weeks → 9 days
  • Rebuilt tech support: incident response 3h → 45min (-75%)
  • Built real-time ops dashboard (Grafana / Kibana)
Fintech

Promsvyazbank

2021 – 2022

Head of Digital Product Development

Major Russian bank

  • Built semi-automated contract renewal: CRM trigger → bot → one-click confirmation
  • Integrated with contact centre (auto-dial, IVR scripts), cut processing 2 days → 4 hours
  • Launched auto-insurance campaign: +18% conversion to policies
Insurance

Rosgosstrakh

2019 – 2021

Product Manager

My_Service digital insurance ecosystem

  • Product and project manager for digital insurance platform serving 50k+ monthly users
  • Led product strategy and delivery for auto-services vertical

Beyond Work

Not Just Spreadsheets

Building products is what I do. Here's what keeps me sharp outside of it.

🏄 Surfing
🪎 Kitesurfing
⛰️ Hiking
🛫 Travel
🎿 Snowboarding
🧗 Bouldering
@pskuzko

Get in Touch

Let's Talk

Open to full-time remote or relocation to EU. AI Product Manager roles in fintech.

Email Me LinkedIn

Or download my CV directly.