AI automation that works for your business
VorkLab is an AI engineering practice led by Valentin Shapovalov, with 15+ years in IT. We help business owners and teams build AI assistants and automate workflows, drawing on production experience in marketplaces, medtech, and e-commerce.
Tell us about your challenge. We will suggest a practical next step, with no obligation.
Services and pricing
All prices in USD. Final scope and quote agreed before work begins.
Start here
AI consultation and audit
A 60-minute call and a written roadmap: what to implement and in what order.
One-time
Consultation & written roadmap
1 week
- 60-minute discovery call
- A structured assessment of your needs
- Written roadmap with priorities
- Recommended next step
AI assistant
Your own AI assistant
A Telegram bot or OpenWebUI assistant with retrieval over your documents and knowledge base.
From
+ ~$50/month hosting
3–7 days to set up
- Deployment on your server or managed infrastructure
- Answers grounded in your documents and knowledge base
- Two 30-minute onboarding calls
- One month of support and prompt updates
Workflow automation
Automate one workflow
An FAQ bot, lead processing, a team assistant, or reporting: one workflow delivered end to end.
From
Per workflow
2–4 weeks
- Discovery → MVP → iteration → handover
- One agreed workflow: support, leads, content, reporting, or booking
- Before-and-after metrics and A/B testing
- Documentation and team handover
What we automate
FAQ and customer support
Answer common customer questions using your knowledge base and hand complex requests to a person. Add multilingual support where needed.
From prior engineering experience at a marketplace with 20M+ buyers
Appointment booking and reminders
Handle voice and text enquiries, qualify requests, book or reschedule appointments, and send confirmations and reminders.
Prior engineering work; client details confidential
Incoming lead processing
Classify incoming requests, populate CRM fields, and route each lead to the right person.
A knowledge assistant for your team
Give your team an assistant that retrieves answers from your documents, including Notion, Confluence, and Google Drive.
Defined scope and fixed setup price
Product descriptions and content
Generate product descriptions, catalogue content, and SEO pages using templates and your brand voice.
Segmentation and personalisation
Segment customers, predict lifetime value and churn, and detect anomalies. Connect the results to CRM and retention workflows.
From prior ML engineering work at a game studio
Free guide + reusable prompt
Plan your first automation.
Choose one task, define a good result, and decide how to test it. Our practical guide helps you turn an idea into a clear implementation plan with your AI assistant.
Read the guideRead online, copy the prompt, or download it. No sign-up.
Leave with a plan that covers
- A clear scope and a result you can check
- Human decisions, spending limits, and recovery
- A small trial before a live rollout
Engineering experience
Selected work from Valentin Shapovalov’s engineering career, including roles before VorkLab. These projects show the experience behind the practice.
Customer support AI for a major marketplace
AI/ML Engineer · 20M+ buyers, 5 languages
- Multilingual RAG + agentic workflows (PL / CZ / SK / HU / EN)
- Proof of concept to production in 26 weeks
- ~45% auto-resolution, CSAT ~76%, p95 ≤ 350ms
- Hybrid retrieval over 40M+ chunks; 5× reduction in hallucinations using a critic and reranker
Voice and text AI assistants for a medtech service
AI Engineer · production traffic at clinics
- Handling and qualifying incoming voice and text enquiries
- Booking, rescheduling, cancellations, confirmations, and reminders
- LangGraph and voice tools (Eleven / Silero / Yandex), with domain-specific retrieval
- Production evaluations and guardrails to detect unsupported answers
Recommendation systems for a major e-commerce retailer
AI/ML Engineer · production, AB-tested
- Similar-items (HNSW + ALS): +8% CTR
- Complementary-items (co-purchase HNSW / ALS): +4.6% GMV
- Learning-to-rank (CatBoost / YetiRank): +17% MAP
Lifetime value, churn, anomaly detection, and segmentation for a game studio
ML Engineer · production CRM impact
- Production lifetime-value model: +19% marketing campaign ROI
- Churn prediction (CatBoost / SGB) + anomaly detection (IsolationForest)
- Player clustering (DBSCAN) for personalised CRM workflows
About
Valentin Shapovalov
AI/ML Engineer · Founder of VorkLab · 15+ years in IT
I lead VorkLab and build AI systems from discovery to deployment: text and voice agents, retrieval, and evaluations. My engineering experience spans marketplaces, medtech, and e-commerce. I share practical lessons on @focus_ops and publish components and templates on GitHub.
Tell us what you want to improve
In a free 30-minute call, we will review your current process and discuss where AI could help. You will leave with a clear next step, with no obligation.
Book a 30-minute discovery call