1
No clear path after senior
“I've made it to Senior – I don't understand what's next or where to look”
The course gives you a roadmap: what a tech lead does, what an architect does, how they differ and how to get there.
Run 21 · starts 21 September 2026 · free consultation with Pavel Veinik
Sign up → Course Technical Leadership
Senior → Tech Lead → Architect
Foggy after Senior? There is a path.
AI runs through every module – plus three modules entirely on AI
Live online sessions, recorded · Free consultation with Pavel: find out whether this course is right for you and what it will give you
“I stopped selling ideas and started speaking the same language as the business” – Ilya Skumin, Tech Lead, Fintech
“In depth and practical value, a head above TOGAF and SEI” – Ilya Chakun, Software Engineer, Zürich
“Architecture decisions stopped being a black box” – Alexey Terentyev, Senior SWE → System Architect
400+graduates
21course run
12+12theory + practice
∞access to materials
01
Who the course is for – and what you take away
Find your level. The course takes you further – Middle+ → Senior → Tech Lead → Architect.
Middle+
You write code with confidence, but you have never put together a large system end to end. AI is everywhere, and how to fit it into the architecture is unclear.
- Discover distributed systems architecture: what large systems are built from
- Get a template for solving architecture problems – the foundation everything else rests on
- +AI learn to design AI systems – not separate calls, but the whole system
Senior
You made it to senior – and beyond that it is fog. You don't want to go into management, and a technical track seems not to exist.
- Systematize your experience: it stops being a collection of separate cases
- Learn to choose databases, queues and caches by criteria, not by habit
- Understand where to go after senior on the technical track
- +AI learn to calculate what AI costs – tokens, GPUs, inference
Tech Lead / Architect
You make decisions but can't justify them – you know intuitively. Your tasks are above your level, and there is no one inside to ask.
- Learn to justify a decision and speak the business's language: risk, money, deadlines
- Master auditing an unfamiliar system: recovering decisions, ADR, risk map
- Figure out when to go monolith and when microservices – and how it relates to how the team is organized
- +AI audit and red-team your own design together with AI
02
Who it fits – and who it definitely doesn't
This course is not for juniors. Without experience on complex production projects, the material won't land.
✕
Not your course if
- You have less than 3 years of commercial development experience
- You have never worked in a team on a real production project
- You are looking for a course on a specific stack or language
- You need a certificate for your CV, not a way of thinking
✓
Your course if
- 3+ years of experience, Middle+ level or higher
- You feel a ceiling and don't see a clear path forward
- You want a technical growth track, not a management one
- You are ready to spend 10 hours a week for 3 months
03
9 problems the course solves
What the course solves, based on in-depth interviews with graduates.
2
Architecture decisions by gut feeling
You pick a database, queue or cache out of habit or from Stack Overflow.
After the course: 20 database criteria, 13 queues, 17 caches. A systematic choice you can explain to the team and the business.
3
You can't justify a decision to the business
Technically ready. But with directors, it's a different language.
Module 12: architect communications – how to explain decisions, risk and AI non-determinism to the business.
4
Unclear what AI changes in architecture work
Agents, RAG and vector databases everywhere. Where the architecture ends and the hype begins is unclear.
AI runs through every module – plus three modules on AI: components and selection criteria, agents and MCP, RAG and serving, designing and auditing architecture with AI. Plus what few people have really figured out: token costs and GPU economics, evals for managing inference quality, model drift, prompt injection and data leaks.
5
Knowledge without a system
You've read the books, watched the videos, worked on projects. It's all in your head, but it doesn't add up to one picture.
The course systematizes what you already have.
6
Impostor syndrome before the next role
“It felt like I wasn't ready yet.” Or the opposite: “everything has become so easy that I start wondering – maybe I've just stagnated.”
A free consultation with Pavel before you pay and reviews of your own tasks during the course give you an outside assessment of your level – instead of a self-assessment.
7
No architecture practice – only theory
Lectures into thin air don't help.
Workshops: we design a URL shortener and work through participants' real tasks. Practice with feedback.
8
Tasks above your level – no support inside
You've been thrown onto a complex project. There's no one at the right level inside the company.
Access to a practicing architect's way of thinking and to the alumni community.
9
Monolith or microservices – unclear when to pick which
No systematic answer.
Module 12: microservices, monoliths and organizations – when each is chosen and how it relates to how the team is organized.
04
Who teaches the course: Pavel Veinik
Pavel Veinik
“A developer can't grow without understanding the interests of the business”
05
Watch before you buy
Pavel has been publishing open content since 2016. Take a look and decide for yourself.
YouTube – latest episodes
Pavel's open lectures and breakdowns
- “How to manage tech debt in practice” – Tech Debt Management
- “How is the architect's role changing in the AI era?” – round table
- “How modern RDBMSs work, using PostgreSQL as an example”
Telegram – articles and matrices
The Hard&Soft Skills channel
- Articles on architecture and systems thinking
- Tech lead and architect competency matrix
- 8 years of content – plenty to read
06
Read before you buy
Pavel has been writing since 2016. Articles on architecture, engineer growth and systems thinking. The same level as the course – just shorter. The articles are in Russian.
25 AUG '26 · 19:00 (GMT+3) · ~1.5 hours
Open meetup: Technical Leadership
The Senior glass ceiling – where it comes from and how to break through? How to grow after senior in the AI era – the technical track, no fluff.
With Pavel Veinik, ex-Architect at Miro and EPAM. In Russian, recorded.
07
Course program: 12 modules plus a bonus one
12 modules plus a bonus one. 5 parts. Not “what we'll study”, but “what you'll be able to do afterwards”. AI runs through every module – the +AI tag shows exactly where. Plus three modules entirely on AI.
Part I – Technical context
Module 1: What makes a good tech lead
- What makes a good tech lead +AI: non-determinism as a new kind of complexity in an architect's work
- Non-functional requirements +AI: NFRs for AI, token costs, security, non-determinism
- Participants' Q&A
Module 2: Limits of the possible, quality criteria, laws of distributed systems
- Constraints and principles +AI: non-determinism as a class of complexity, GPU constraints
- A template for solving architecture problems +AI: the method applies to AI systems too
- Problem statements: URL shortener, visit counter, chat +AI: RAG / agent / inference problem statements
- Designing the architecture of a URL shortener
- Designing the architecture of a visit counter
- Participants' Q&A
- What “a lot” and “a little” mean +AI: calibrating scale – tokens, context, embeddings, GPUs
Module 3: Distributed systems patterns. Integration patterns
- Distributed systems patterns +AI: LLM resilience – fallback to other models, response caching
- Integration patterns +AI: LLM/agent as a service, MCP as a standard
Part II – System components
Module 4: Systematizing architecture components. Databases
- Systematizing architecture components +AI: AI components in the taxonomy
- Systematizing databases: part 1
- Systematizing databases: part 2 +AI: vector databases
- Systematizing databases: part 3
- Systematizing databases: part 4 – selection criteria +AI: vector search / embeddings criterion
- The psychology of an architect +AI: decisions about unfamiliar AI technologies
Module 5: Systematizing message queues and caches
- Systematizing message queues +AI: queues for async / batch inference, token streaming
- Systematizing caches +AI: semantic response cache, prompt caching – saving tokens
Module 6: Systematizing load balancers and file storage
- Systematizing load balancers +AI: routing between models/providers, GPU-aware
- Load balancer layers
- Systematizing file storage +AI: storing model weights, datasets, embeddings
Part III – AI architecture and designing with AI
Module 7: Systematizing AI components
- Foundation models: classes (LLM / LMM / SLM) and selection criteria
- Vector databases and embedding stores: selection criteria
- Orchestrators and agent frameworks: selection criteria
- Inference and serving solutions: selection criteria (self-hosted vs API, GPU economics)
- RAG infrastructure: components and selection criteria
Module 8: Advanced AI architecture
- Agent types and orchestration; multi-agent architectures; when an agent is overengineering
- Model Context Protocol (MCP)
- The evolution of RAG patterns
- Serving layer: architecture, API Gateway, routing (LiteLLM / Kong); caching, rate limiting, graceful degradation
- Observability for AI systems
Module 9: Designing and auditing architecture with AInew
- AI in the architecture decision loop: where it amplifies (breadth of options, drafting, challenging), where it predictably fails (cost, latency, vendor limits). Responsibility for the decision is not delegated
- A set of criteria as an executable rubric: generating and narrowing the solution space – a table with calculations instead of opinions, explicit “don't know” marks
- AI as an opponent: red-teaming your own design against NFRs – where the solution breaks at x10
- Auditing an existing system: reverse-engineering C4 from the repository, recovering undocumented decisions, ADRs, a risk map, fitness functions
- Accepting someone else's or an AI-generated solution: a review rubric, verification discipline, signs of the plausibly wrong
- Homework: auditing a real system by protocol – report + ADR + risk map
Part IV – Quality, deployment, security
Module 10: Quality management, deployment and monitoring
- Quality management +AI: evals for managing inference quality, testing the non-deterministic, chaos for AI
- Migrations and zero-downtime deployment +AI: versioning, canary and rollback of models/prompts
- Monitoring tools and approaches +AI: LLM observability, drift, cost monitoring
Module 11: Architecture securitynew
- Threat modeling and attack surface
- Authentication and authorization: OAuth 2.0 / OIDC, JWT, RBAC/ABAC
- Secrets, encryption and network security
- AI system security: prompt injection, data/PII leaks, agent security, model supply chain
Part V – Organization, communication, career
Module 12: Microservices, organizations, communication and career
- Microservices, monoliths and organizations +AI: agent architectures as a distributed system, build-vs-buy for LLMs
- Architect communications +AI: how to explain AI risk and non-determinism to the business
- Career: tech lead and beyond +AI: how AI affects an architect's career
Bonus module 13: Extras
- Systematizing distributed data processing +AI: embedding pipelines, batch inference, feature store
- Systematizing ORM frameworks
- Systems engineering fundamentals +AI: AI systems as an object of systems engineering
- How Apache Kafka works
- How the Raft consensus algorithm works
- Distributing a system across data centers +AI: serving models by region, data residency
- A conversation with a guest expert (for the communication part)
- Tools for CQRS and Event Sourcing
08
Course price: three packages
Three packages – choose by your goals and budget. You fill the consultation hours yourself: not all topics will fit, so you and Pavel put the plan together at the first meeting.
Standard
$2,600
or 3 payments × $940
- Theory: live lectures, access to recordings, notes
- Practice and homework reviews with Pavel's feedback
- Certificate in RU + ENG
- Access to the alumni community
- All materials – forever
Career
$3,500
or 3 payments × $1,250
- Everything in Standard
- 3 hours one-on-one with Pavel – you choose the topics
Skills diagram assessment
- a completed diagram showing gaps against your target role
- a plan to close the gaps: what the course covers, what to pick up separately
Your path and move within your company
- a scope expansion map: which decisions and areas of responsibility to take on in your current role, and in what order
- a list of evidence for the promotion conversation – design docs, results in numbers
Career path on the market
- a target role and 2–3 market segments that hire this profile, with the order to approach them
- a step-by-step transition plan with deadlines and checkpoints
A CV for your target role
- a rewritten CV: role framed by the job's function, experience framed by scope of responsibility
- 2–3 cases from your experience, packaged for interviews: decision → alternatives → result
System design interview rehearsal
- a review of the run against a hiring rubric: what the interviewer counted, what they didn't and why
- a list of weak spots with a preparation plan before the real interview
Architect
$4,000
or 3 payments × $1,420
- Everything in Career
- 5 hours instead of three – these topics or any from Career
An audit of your production system
- a risk map: what breaks first, under what conditions, and the cost of failure
- a prioritized fix plan: quick wins separated from structural changes
- recovered ADRs – a log of the system's key decisions with their rationale
Design sessions on your task
- a solution designed with the 7-step method and a set of criteria
- a design doc for the solution: trade-offs, rejected alternatives, limits of applicability
Justifying a decision in writing and in person
- a one-pager for the business and the client: decision, risks, cost – without technical jargon
- a rehearsal of defending the decision (pre-sales, architecture board) with a review of your arguments
AI system architecture review
- a verdict on the system: bottlenecks in token cost, latency, answer quality
- a map of AI-specific risks: non-determinism, prompt injection, PII leaks, quality drift
- recommendations on evals and monitoring – how to measure quality after the review
Review of how you use AI in design
- a review of your practice: where AI measurably speeds you up and where it creates risk
- a protocol for working with AI on your tasks: red-teaming your own solution, accepting AI proposals
- a set of rubric prompts for checking solutions proposed by AI
The course gives you the method and group reviews of study tasks. A consultation applies the same method to your real work case, one-on-one with Pavel.
Questions about payment – which country to pay from, installments, company payment and invoices? Message our manager on Telegram →
⚑ 400+ graduates have already completed the course · Run 21
09
What graduates say
“The most noticeable change after the course is confidence in myself and my decisions.”
I had been following the Technical Leadership course for a long time but kept putting it off: it seemed I wouldn't cope with the amount of material and simply wouldn't be accepted.
“Hard&Soft Skills are the only courses after which you really grow both technically and as a leader.”
I went all the way from the TechLead course to the CTO course and can compare them with programs from other big players on the market.
“I stopped ‘selling ideas’ and started speaking the same language as the business.”
For me, this course was useful not just as a source of new knowledge – it helped me turn the experience I already had into a system and put it to use right away.
“I came to systematize 18 years of experience and gain confidence for pre-sales of architecture solutions – and got exactly what I wanted.”
I came to the course after more than 18 years in IT – to systematize my knowledge and gain confidence for pre-sales of architecture solutions, and got exactly what I wanted. The format of “you read the material yourself whenever it suits you, and…
“A great course that helped me put my accumulated experience in order and learn something new.”
A great course that helped me put my accumulated experience in order and learn something new. A special plus for the AI modules – the breakdown of agent interaction architecture and core principles was very relevant.
“The template really clicked: even when I don't apply it literally, it plays out in my head and shapes how I justify decisions.”
The course met my expectations in exactly what mattered to me – a structured approach to an architect's work, from the client's wishes to the final solution. The template really clicked: even when I don't apply it…
“After the course, architecture decisions stopped being a ‘black box’: I started to understand why a particular choice was made and what the alternatives are.”
The main value of the course for me is systematized architecture knowledge and understanding the reasons behind every decision.
“The course is built to show that technology is not the most important thing.”
The course exceeded my expectations. I came for a deeper dive into system design, but the first parts opened my eyes to things I hadn't paid attention to – gathering requirements, arguing for approaches; and I've already started putting a lot into practice…
10
Community
A Telegram channel, a cohort chat and an alumni community for life.
A living community – Telegram
An open channel with content. A private chat for participants of the current cohort. Discussions, questions and reviews of architecture tasks while you study.
Subscribe to the channel →Alumni chat – for life
A private chat for graduates of all cohorts. Access for life. Ask a question about a real project – people who have walked the same path answer.
“I learn about things I would never have heard of otherwise” – graduate, Tech Lead, Fintech
News subscription
Announcements of new cohorts, meetups and open sessions. No more than once a month. No spam.
Subscribe to news →11
FAQ
What's new in this run?
Two entirely new modules. The first is architecture security: threat modeling, OAuth 2.0 / OIDC, JWT, RBAC and ABAC, secrets and encryption, and separately – AI system security: prompt injection, data and PII leaks, agent security, model supply chain. The second is designing and auditing architecture with AI: AI as an opponent to your design, auditing an unfamiliar system by recovering its decisions and building a risk map, accepting a solution generated by AI. Three modules on AI in total, and AI itself runs through every module. We also cover operations: evals for managing inference quality, rolling back models and prompts, drift, cost monitoring.
Is this a course about AI?
No. This is a course about distributed systems architecture and the tech lead role – AI is an element in it, not the subject. But it is a cross-cutting element: non-determinism, token costs, vector databases, semantic caching and routing between models are covered inside the regular architecture modules. Plus two modules entirely on AI architecture, a module on designing and auditing architecture with AI, and a module on architecture security that covers AI system security separately. If you want AI to write code, that's our other course, AI-Driven Development. Here AI takes part in the architecture decision and the audit, while a human is responsible for the decision.
Is my level enough? I'm not an architect
The course is designed for senior engineers and those close to it: you need at least 3 years of commercial development experience and work on real production projects. You can find out whether it's your level before paying – at a free consultation with Pavel. It doesn't commit you to anything: if the course isn't right for you, Pavel will tell you so and give you individual recommendations.
How is this different from the system design materials freely available online?
You can read material anywhere. On the course you get what materials don't give you: a template for solving architecture problems that you work with from the first session, reviews of your own tasks, and personal feedback from a practicing architect on every homework assignment.
How much time does it take per week?
10 hours. You read the material yourself whenever it suits you, and in live sessions it is reviewed and discussed – twice a week, theory and practice, plus homework. The course runs for 3 months.
What if I miss a session?
All sessions are recorded. Access to recordings and materials stays with you forever, with no time limit.
When do sessions take place?
Twice a week: theory and practice. Participants choose the specific days and times together by voting at the first organizational meeting.
Can I pay in installments?
Yes. Any package can be paid in three equal payments.
Will there be a certificate?
Yes, an electronic certificate of completion in Russian and English.
How much does the course cost?
Three packages: Standard – $2,600, Career – $3,500, Architect – $4,000. Any of them can be paid in three equal payments. All packages include lifetime access to materials and the alumni chat.
Who teaches the course?
Pavel Veinik – architect at Miro and EPAM, CTO at AmadoAd Ltd., SplitMetrics and Leverice, founder of Hard&Soft Skills, ITStart and AmadoAd Ltd. A developer since 2003, teaching since 2008, and since 2018 working only with senior engineers and architects. Has run more than 100 meetups and conferences.
What will I be able to do after the course?
Design distributed systems using a template for solving architecture problems. Choose databases, queues, caches and load balancers by criteria, not by habit. Design AI components: RAG, agents, inference and serving. Cover architecture security, including prompt injection and data leaks. Justify architecture decisions to the business.
How does a tech lead differ from an architect?
They are different roles, and how they differ is the topic of the course's first lesson and of modules 1 and 12. You can watch the first lesson before paying: leave a request and we'll email you a link.
What language is the course in?
Sessions and materials are in Russian. The certificate is issued in Russian and English.
12
Run 22 – coming soon
Meanwhile, book a consultation with Pavel
- 12 theory + 12 practice sessions, recorded – yours forever
- A template for solving architecture problems
- Course materials with no time limit
- A private chat for the current cohort
- Alumni chat for graduates of all cohorts – for life
- Reviews of participants' real architecture tasks
- Three modules on AI: components and selection criteria, agents and MCP, RAG and serving, designing and auditing architecture with AI
- A new security module: threat modeling, OAuth 2.0 / OIDC, RBAC and ABAC, prompt injection, data and PII leaks
- Operations: evals for non-deterministic systems, zero-downtime, rolling back models and prompts, AI cost monitoring
Format: you read the material yourself whenever it suits you, and live sessions review it – twice a week, online, recorded. 3 months. 10 hours a week. Sessions are in Russian.
Run 21 has already started – it's no longer possible to join. We'll announce the start of Run 22 soon. Don't wait for the announcement: book a consultation now and find out whether this course is right for you.
The consultation with Pavel is free. It doesn't commit you to anything: it's there so you can understand whether this course is worth it for you and whether it will meet your learning goals. Pavel will help you see your strengths and how you can grow in three months. If the course isn't right for you, he'll tell you so during the consultation and give you individual recommendations.
Still have questions? Message our manager on Telegram →
Hard&Soft Skills · made by engineers for engineers · updated 06 / 2026
Получить пробный бесплатный урок.
После отправки данной регистрационной формы на указанную почту придет ссылка на бесплатный демо-урок.
It's totally free