Fine-tuning or RAG? The expensive mistake happens before training
Fine-tuning vs RAG for business: diagnose the failure, compare costs and test the result before paying to train a model.
Read the articleDANIIL SHAKHOVSKIY · AI PRODUCT LEAD & HEAD OF PRODUCT
Good products.
Better workdays.
I’m Daniil Shakhovskiy, an AI product lead and hands-on builder. I turn complicated problems into useful products — and make the working day a little more human.
Welcome to my personal officeBased in Czechia. Open to teams worldwide. Good work starts with a real conversation.
Product thinking. Hands-on building. People first.ONE HUMAN. SEVERAL DEPARTMENTS.
NOTES FROM THE OFFICE
Real work.
Real questions.
AI that helps. Hiring that makes sense. Teams that want to stay. A little reading before the next big decision.
Open the article feedThree questions on the desk right now.
THE EDITOR’S PICKS / 01 / 02 / 03
Public data is not permission: six official sources AI teams should check
Where can AI teams get data officially? Six sources, their useful roles and the provenance questions to settle before training or building RAG.
Read the article
The moat is a rented API. What should an investor ask next?
AI startup due diligence beyond the demo: examine customer dependence, data rights, workflow ownership, supplier risk and the cost of a completed outcome.
Read the articleKEEP SCROLLING
More from the office.
Find the question your team is asking.
Or the one nobody has said out loud yet.
50 articles. Pick what sounds familiar.
Meet the person behind the notes
Six AI companies worth studying before you copy another demo
A practical AI company watchlist for founders and product teams: what to study at OpenAI, Anthropic, Google DeepMind, Mistral, Hugging Face and NVIDIA.
Read the article
95% accuracy. A very expensive 5%.
AI evaluation for business teams: why average accuracy hides costly failures, and how to test severity, coverage, review effort and real customer outcomes.
Read the article
An AI marking its own homework: the synthetic data problem
Synthetic data can help AI development, but who checks it? A business guide to model collapse, independent tests, provenance and useful generated examples.
Read the article
Your fine-tuned model has an expiry date. Who owns the migration?
Fine-tuned model migration: current OpenAI shutdown dates, dependency checks and a practical plan to preserve quality when a base model is retired.
Read the article
The enterprise deal died in a security questionnaire
Enterprise AI sales stall when data handling and ownership are unclear. A practical guide to security questionnaires, procurement evidence and pilot scope.
Read the article
Would anyone renew if the AI stopped being new?
AI SaaS retention beyond launch metrics: measure repeat customer outcomes, distinguish users from buyers and prepare a renewal conversation with evidence.
Read the article
“We own the model.” Great. Who is on call?
Open-weight models vs hosted APIs for business: compare rights, staffing, reliability, utilisation and cost per accepted outcome before choosing deployment.
Read the article
AI integration costs: what your quote should actually include
What should AI integration cost? A worked budget, operating-cost checklist and questions to ask before approving an implementation quote.
Read the article
AI agent or workflow automation? Stop hiring a robot for a checklist
Do you need an AI agent, an AI-assisted workflow or ordinary automation? A practical decision framework for business owners and product teams.
Read the article
Your company chatbot knows everything. Including what it should not.
Planning an internal AI knowledge assistant? Start with document ownership, permissions, source quality and a practical acceptance test for your team.
Read the article
Hire an AI engineer, an AI consultant or an AI Product Lead?
Hiring for an AI project? Compare an AI engineer, consultant and AI Product Lead by the work you need, with an interview brief and evidence to request.
Read the article
Your AI agent has permissions. Who has responsibility?
Who is responsible when an AI agent acts? A practical guide to ownership, permissions, human review and stop rules for founders and product leaders.
Read the article
No junior jobs, no future seniors: the AI hiring trap
AI and entry-level jobs: what recent employment research does and does not show, and how employers can build a junior talent pipeline without busywork.
Read the article
AI workslop: the polished document making your team slower
AI workslop makes colleagues finish the thinking. Learn how to recognise it, measure review costs and set a useful quality standard for AI-assisted work.
Read the article
Your employees already use AI. Your policy missed the meeting.
Employees using unapproved AI tools? Find the unmet need, protect sensitive data and build an approved route people can actually use.
Read the article
Skills-based hiring: stop interviewing the CV
A practical skills-based hiring process: define outcomes, use fair work samples, set an AI policy and score evidence instead of CV polish.
Read the article
Is your candidate real? Hiring trust in the age of AI interviews
How to respond to AI candidate fraud and deepfake interview concerns with fair assessment, proportionate verification and a consistent hiring process.
Read the article
Ghost jobs: why your careers page is losing candidates' trust
Ghost jobs and candidate ghosting damage employer trust. Label real vacancies, close stale listings and build a hiring update process that people can believe.
Read the article
Your hiring AI rejected them. Can anyone explain why?
Can your hiring team explain an AI rejection? A practical review of screening criteria, meaningful human oversight, candidate routes and EU AI Act timing.
Read the article
Competitive salary is not a number: pay transparency in 2026
EU pay transparency in 2026: meaningful salary ranges, clear levels and better hiring conversations, with links to the current Commission guidance.
Read the article
Will AI take my job? Your team needs more than reassurance
How managers can discuss AI job anxiety honestly, explain what is changing and create practical support without making promises they cannot keep.
Read the article
AI saved an hour. Your boss added three meetings.
Why AI time savings can turn into more work. A practical workload experiment covering review costs, new demand, capacity and sustainable team productivity.
Read the article
After the layoffs, the work stayed. Now what?
After layoffs, remaining teams need a realistic workload reset. Map obligations, stop work, rebuild ownership and communicate what can actually be delivered.
Read the article
Your next great hire might already work here
Before recruiting externally, check internal talent. Build a fair mobility process with visible opportunities, work evidence, transition plans and manager support.
Read the article
The best employee is not always the one nearest your desk
Proximity bias can distort performance and promotion decisions. Learn how hybrid teams can assess contribution, share opportunities and use office time well.
Read the article
Forward deployed engineer: the person between the AI demo and real work
What does a forward deployed engineer do? Compare FDE, product engineer and consultant roles, assess the mandate and define value beyond an AI demo.
Read the article
Your company cannot scale if every decision needs you
Is the founder becoming the bottleneck? Diagnose approval delays, delegate decisions safely and build a company that can move without waiting for you.
Read the article
Hiring an AI Product Lead: the CV tells you less than you think
How to hire an AI Product Lead: assess business judgment, AI evaluation, delivery and team alignment with a practical interview scorecard.
Read the article
Your best employee did not leave for the free coffee
Why do good employees quit? A practical retention guide for founders and HR: diagnose pay, workload, trust and career friction before the resignation.
Read the article
You paid for AI training. Why is nobody using it?
Why employees ignore AI after training, and how HR and managers can turn a workshop into safe, useful habits built around actual work.
Read the article
Your team is busy. Why is nothing getting finished?
A busy team with poor results may have too much unfinished work, unclear acceptance or slow decisions. Diagnose the flow before demanding more effort.
Read the article
The star performer everyone dreads: is the output worth it?
How to manage a high performer whose behaviour harms the team: assess the whole contribution, investigate fairly and set clear working expectations.
Read the article
Hiring a Head of Product? Stop shopping for a miracle
How to hire a Head of Product who fits your business: define the mandate, test judgement, agree decision rights and build a useful first 90 days.
Read the article
Your AI demo works. Your business still does not feel it.
An AI pilot can impress management and still fail at real work. Diagnose missing ownership, weak evaluation and adoption before paying to scale.
Read the article
You hired the right person. Your onboarding is losing them.
Employee onboarding failing? Help new hires become effective with clear responsibilities, useful access, real feedback and a practical first-month plan.
Read the article
Before you hire five more people, fix these five bottlenecks
Before adding headcount, check five business bottlenecks: decisions, rework, handoffs, manual work and priorities. Identify when hiring is genuinely needed.
Read the article
Your team is resisting change. What are they protecting?
Why employees resist change and what managers can do: uncover real concerns, test assumptions and make AI or process adoption workable.
Read the article
AI automation ROI: the costs your demo conveniently forgot
Calculate AI automation ROI with implementation, review, maintenance and real benefit capture. A fictional worked example separates cash savings from capacity.
Read the article
Your roadmap is full. Your strategy may be empty.
Too many priorities on your product roadmap? Connect commitments to customer problems, make tradeoffs explicit and replace feature overload with a workable strategy.
Read the article
Need a product leader, but not another full-time salary?
When a fractional product lead makes sense, when full-time leadership is necessary, and how founders can define scope before hiring external help.
Read the article
Sales promised it. Product never agreed. Now what?
How founders and product leaders can handle unapproved sales promises, agree commitments and stop customer handoffs turning into internal firefights.
Read the article
10 difficult conversations to have when you want a promotion
How to talk to your manager about a promotion: pay, proof, vague feedback, budget objections and a clear next step. Practical phrases for ten hard conversations.
Read the article
8 types of employees who help your business scale
Eight ways employees help a business grow, from finishing work to teaching others, with practical questions for managers building stronger teams.
Read the article
Remote, hybrid or office: which way of working fits your team?
Compare remote, hybrid and office work using research, team needs and a practical cost model. A guide to choosing and testing the right fit.
Read the article
AI is an amplifier. You still need a point of view.
A human guide to working with AI: broaden your skills, experiment with purpose, and delegate tasks without handing over your judgment.
Read the article
10 difficult people at work: better ways to talk to them
Ten familiar workplace behaviours, from the moving goalposts to the permanent emergency, with practical phrases and boundaries for better conversations.
Read the article
AI adoption for businesses: a practical, people-first guide
How to introduce AI into a business: choose a useful workflow, involve your team, run a measured pilot, and build training around real work.
Read the articleNo meeting on that topic. Yet.
Try a broader search, or start again with all the articles.
Recognise something in your own team?
I help turn complicated product, AI and people problems into a useful next step.
01 / THE PERSONAL OFFICE
Est. one person ago.Meet the departments.
They look familiar.
Seven departments. One Daniil.
Different hats for different problems.
Yes, the internal meetings are complicated.
IMPORTANT HR UPDATESame person.
Same person.
Less overhead.
These are playful alter egos, not seven employees. You work directly with me.
Finally, a meeting you can enjoy.
COMPANIES, VENTURES & CLIENT WORK
I’ve built businesses of my own and worked with teams across enterprise software, hospitality, consumer technology, AI research, corporate training and venture investment.
- RobosizeMEHead of Product & Forward Deployed Engineer
- DeloitteExternal AI research consultant
- Veeam SoftwareProduct analysis & employer brand
- AbridgeExternal ML research assistant
- Coders Lab CZ & SKCorporate AI trainer
- Initio CapitalVenture Partner
- Lexicon LabsCo-founder & AI automation agency lead
- My HeirloomCo-founder, CTO & board advisor
- TAMADA GroupFounder & business owner
- InfluenceUPFounder & product owner
- InfluenceAR NFTFounder & product owner
26 projects across Europe. Web, mobile, design and research projects delivered through TAMADA Group, from client discovery and scope to execution.
Two independent products, two exits. I founded and built InfluenceUP and InfluenceAR NFT, taking both through launch and successful exits.
This experience includes employment, external collaborations and businesses I founded or co-founded. My agency and consulting work also includes clients not named individually here.
Explore the roles and projects02 / FIND YOUR SITUATION
Straight to the useful bit.When to bring
me into the room.
Pick the situation that sounds familiar.
See how I would start, what we would work towards and who should join the conversation.
PRODUCT STRATEGY & DISCOVERY
Agree on what is worth building.
For a team pulled between customer requests, business goals and technical constraints. I connect customer and market research with product decisions, then help the team commit to a practical direction.
- WHERE I START
- Talk to customers and stakeholders, review the product and map the decisions that are blocking progress.
- WHAT WE WORK TOWARDS
- A prioritised opportunity map, a product roadmap and a clear next step with an owner.
- BRING ME IN WITH
- Founder, CEO, CPO or product lead
RELEVANT EXPERIENCECompany-wide product direction at RobosizeME; AI research and recommendations for Deloitte.
Let’s clarify the direction ↗AI OPPORTUNITIES & WORKING PILOTS
Turn AI interest into a useful first project.
For a company with plenty of AI ideas and no obvious starting point. I assess the real workflows, users, data and adoption risks, then help build and test a focused pilot.
- WHERE I START
- Review a real workflow with its owner and compare the effort, usefulness and limits of possible AI use cases.
- WHAT WE WORK TOWARDS
- A prioritised use case, a working pilot and a practical way to judge whether it helps.
- BRING ME IN WITH
- CEO, CTO, innovation lead or operations director
RELEVANT EXPERIENCEAI technology research at Deloitte; GenAI products and automation at Lexicon Labs.
Let’s find your first useful AI project ↗WORKFLOW AUTOMATION & KNOWLEDGE SYSTEMS
Give repetitive work a better home.
For teams moving information between tools, searching through documents or repeating manual handovers. I build integrations, AI agents and retrieval-based knowledge systems around the actual workflow.
- WHERE I START
- Follow the task end to end, identify its exceptions and agree on where a human should stay in control.
- WHAT WE WORK TOWARDS
- A tested workflow or knowledge-system pilot, with clear handovers and a plan for everyday use.
- BRING ME IN WITH
- COO, operations manager, CTO or department lead
RELEVANT EXPERIENCEHospitality automation at RobosizeME; document automation and enterprise knowledge systems at Lexicon Labs.
Let’s look at the repetitive bit ↗FROM DISCOVERY TO LAUNCH
Get from the idea to something people can use.
For founders or product teams moving towards a first release, product-market fit or a new market. I connect discovery, scope, product architecture, delivery and go-to-market decisions.
- WHERE I START
- Define the user, the problem and the smallest useful release. Make the trade-offs visible before the team starts building.
- WHAT WE WORK TOWARDS
- A scoped first version, delivery priorities and a feedback loop with real users.
- BRING ME IN WITH
- Founder, CPO, CTO or new-ventures lead
RELEVANT EXPERIENCETwo product exits with InfluenceUP and InfluenceAR NFT; 26 delivered agency projects at TAMADA.
Let’s get the product moving ↗CORPORATE TRAINING & ADOPTION
Make the new skills useful the next morning.
For teams that need practical confidence, not another AI presentation. I design hands-on sessions around the work people actually do, explain the limits and give everyone room to ask questions.
- WHERE I START
- Understand the team’s roles and current skills, then choose relevant tasks to practise together.
- WHAT WE WORK TOWARDS
- Role-relevant exercises, reusable workflows and clear guidance for applying AI at work.
- BRING ME IN WITH
- HR, L&D, people lead or a department manager
RELEVANT EXPERIENCEDesigned and delivered corporate AI training for Coders Lab CZ & SK.
Let’s talk about your team ↗PROJECT RECOVERY & TEAM ALIGNMENT
Make the next move clear again.
For a project with unclear ownership, changing scope or difficult handovers. I listen, bring the constraints into the open and help product, business and engineering agree on a workable next step.
- WHERE I START
- Speak with the people involved, separate facts from assumptions and make the unresolved decisions explicit.
- WHAT WE WORK TOWARDS
- An agreed view of the problem, clearer responsibilities and a short recovery plan.
- BRING ME IN WITH
- CEO, product lead, engineering manager or operations lead
RELEVANT EXPERIENCECross-functional delivery at Lexicon Labs and TAMADA; business-continuity planning at My Heirloom.
Let’s untangle what is stuck ↗PRODUCTISING SERVICES & SELF-SERVE FLOWS
Turn repeated delivery into a repeatable product.
For a business whose useful solution still depends on a person setting everything up. I help identify what can become a product, simplify onboarding and connect delivery knowledge with product design.
- WHERE I START
- Map the setup journey and find the steps that repeat, the exceptions and the points where customers need help.
- WHAT WE WORK TOWARDS
- A productisation roadmap and a first self-serve flow to build and test.
- BRING ME IN WITH
- Founder, CPO, CTO or customer-operations lead
RELEVANT EXPERIENCETurning deployment-heavy automations into self-serve products at RobosizeME.
Let’s make delivery repeatable ↗VENTURE ASSESSMENT & STRATEGIC ADVICE
Connect the product story to the business.
For founders refining their product positioning, or teams evaluating an early-stage opportunity. I bring product, market and technical perspectives to questions about differentiation, growth and investment readiness.
- WHERE I START
- Review the problem, market, product evidence and open assumptions. Find where the story needs stronger support.
- WHAT WE WORK TOWARDS
- A clearer product narrative, an assessment of the main assumptions and questions to resolve next.
- BRING ME IN WITH
- Founder, venture team, investor or board member
RELEVANT EXPERIENCEVenture Partner at Initio Capital; board advisory at My Heirloom; startup accelerator and pitching experience.
Let’s sharpen the product story ↗We agree the scope after understanding your situation. The starting points above are examples of how I can help.
FROM THE QUESTION TO THE WORK
A useful place to start.
FOR HR, RECRUITERS & THE PERSON MAKING THE INTRODUCTION
Get the CVThe short version.
For the right person.
Product strategy, hands-on AI delivery and the human side of adoption — with experience across European and U.S. markets.
AI Product Lead: profile for hiring teams ↗
AI Product Lead /
Head of Product
When you need someone to own discovery, priorities and the roadmap, then stay close to delivery, go-to-market and scale.
Discuss the product role ↗Hands-on AI
product delivery
When business needs must become working AI products, integrations, agents and knowledge systems that people can actually use.
Discuss the delivery challenge ↗Corporate AI training
& team adoption
When people need to understand new tools, practise with their own tasks and build the confidence to use them in everyday work.
Discuss the team’s needs ↗AN INTRODUCTION YOU CAN FORWARD
Daniil Shakhovskiy is an AI product leader with 7+ years across B2B SaaS, enterprise software, startups and consulting in European and U.S. markets. He is currently Head of Product & Forward Deployed Engineer at RobosizeME. His background combines AI product delivery at Lexicon Labs, external AI research for Deloitte and corporate training at Coders Lab. He is worth speaking with when a team needs product direction, hands-on AI delivery and someone who can bring people along.
03 / DANIIL SHAKHOVSKIY — THE EXPERIENCE
The full CVHi, I’m Daniil.
I stay for the
messy middle.
The space between a good idea and something that actually works? That’s where I’m useful.
I’ve built startups, worked with enterprise teams and taught people to use AI in their everyday work. I connect product thinking with hands-on delivery — and listen before I suggest a solution.
I’m based in Olomouc, Czechia, and open to working with teams anywhere in the world. What matters is open communication, honest conversations and people who want to make something useful together. We can work in English, Czech or Russian.
Some of the best work never becomes a case study.
I’m comfortable with that.
7+
years across product,
technology & consulting
26
web, mobile, design & research
projects delivered at TAMADA
2
startup products built
through successful exits
THE COMPLETE EXPERIENCE
Roles, collaborations and independent ventures. Product, research, delivery, teaching and the business around them.
MAY 2026 — NOWRobosizeMEHead of Product & Forward Deployed Engineer+
Hospitality workflow automation · Olomouc, Czechia
- Own company-wide product direction: what to build, in what sequence and why. Translate business and customer needs into strategy and prioritised delivery.
- Build and operationalise in-house AI products and integrations across product, engineering, deployment and customer workflows.
- Turn deployment-heavy automations into self-serve products that customers can connect through the website.
DEC 2023 — FEB 2026Lexicon LabsCo-founder / AI Automation Agency Lead+
AI automation and product studio · Prague, Czechia
- Built a personalised AI nutrition assistant using blood-test data that reached $280K in monthly recurring revenue within eight months.
- Delivered GenAI products for document automation and enterprise knowledge systems using GPT-4, Claude, Llama, RAG pipelines, LangChain and vector databases.
- Led seven engineers and designers from discovery to product-market fit, owning the roadmap, prioritisation, delivery and scale-up.
MAY 2022 — APR 2026My HeirloomCTO & Co-founder → Advisory Board Member+
Consumer technology startup · Chicago, United States
- CTO & Co-founder, May 2022–June 2023: set product and technology direction, established operating foundations and aligned the roadmap and resources with growth objectives.
- Created business-continuity and data-protection plans.
- Advisory Board Member, June 2023–April 2026: advised on growth strategy, innovation and resource allocation.
MAY 2025 — JAN 2026Coders Lab CZ & SKAI Corporate Trainer+
Corporate education · Czechia
- Designed and delivered hands-on corporate AI training around the team’s everyday work.
- Helped people adopt AI tools, automate workflows and apply new skills in day-to-day operations.
JAN 2025 — JAN 2026Initio CapitalVenture Partner+
Venture investment · San Francisco, United States
- Sourced and evaluated startups, assessing product differentiation, market opportunity, team strength and investment readiness.
- Connected high-potential founders with strategic investors.
APR 2021 — JAN 2024TAMADA Group s.r.o.Founder / Business Owner+
Digital agency · Prague, Czechia
- Owned delivery of 26 web, mobile, design and research projects across Europe.
- Coordinated client discovery, project scope, execution and outcomes.
JAN 2022 — OCT 2023DeloitteAI Technology Research Consultant — External+
AI technology research and business applications
- Researched emerging AI technologies and commercial applications across business functions, including capabilities, adoption risks and implementation opportunities.
- Turned technical findings into strategic analysis and practical recommendations for business stakeholders.
JUL 2019 — MAY 2022Independent Product VenturesFounder / Product Owner+
InfluenceUP and InfluenceAR NFT · Prague, Czechia
- Built and launched two startup products through successful exits.
- Completed the Seed Starter and Perry Talents SK accelerator programmes and delivered public investor pitches.
MAY 2021 — OCT 2021AbridgeMachine Learning Research Assistant — External+
Six-month external research contract
- Supported ML research through technology analysis.
- Evaluated machine-learning approaches and their practical applications.
FEB 2018 — JUL 2019Veeam SoftwareProduct Analyst / Employer Brand Specialist+
Enterprise software · Prague, Czechia
- Product Analyst, January–July 2019: translated customer change requests and emerging problems into technical requirements and product documentation.
- Partnered with development teams to protect scope, timelines and budgets.
- Employer Brand Specialist, February 2018–January 2019: researched product and brand opportunities and supported positioning, messaging and event-led brand growth.
PRODUCT & BUSINESS
Discovery · User and market research · Product strategy · Roadmaps · Prioritisation · Product-market fit · Go-to-market · Investor communication
AI & DELIVERY
Generative AI · AI agents · Workflow automation · RAG and knowledge systems · Integrations · GPT-4, Claude and Llama · LangChain · Vector databases
PEOPLE & ORGANISATIONS
Cross-functional leadership · Stakeholder alignment · Corporate AI training · Practical adoption · Business-continuity planning
04 / THE HOMEWORK BEHIND THE HUMOUR
Still learning. Always.A little theory.
A lot of doing.
Technical foundations, startup experience
and teaching that starts with real work.
Applied Informatics
Bachelor’s degree, University of Finance and Administration, Prague.
Thesis: traffic-sign recognition using convolutional neural networks and the GTSRB dataset.
From idea to business
Seed Starter & Perry Talents SK.
Built and launched startup products, refined product direction and pitched to investors.
Skills people can use
Corporate AI training at Coders Lab CZ & SK.
Hands-on sessions around the team’s actual tasks. Clear explanations, practical exercises and room for the “obvious” questions.
SOMETIMES I SAY IT IN PUBLIC, TOO
Forbes AI adoption ↗ČT24 AI in the real world ↗iDNES AI at work ↗Svět hospodářství The future of work ↗Dům & Zahrada AI, skills & work ↗
Selected expert commentary · Articles in Czech
06 / MY DOOR IS OPEN
Tell me the
complicated bit.
You can bring the unfinished idea, the stuck project,
or the thing nobody quite wants to deal with.
We’ll start with a conversation.
No pitch deck needed. Coffee optional.