News

We are launching a new course in our catalogue — LLM Tester: quality assessment of AI systems (with Python basics). The first cohort starts on 6 October 2026, with a second group on 27 October 2026.

As LLMs and generative AI systems move into production, checking their quality and safety has become a distinct specialisation with a low entry barrier. The course suits both newcomers to the field and practitioners — testers, analysts, linguists and developers.

What you will learn

• Python for testing: HTTP requests, working with JSON, calling models via the OpenAI and Anthropic APIs;
• how LLM systems work: tokens, context, non-determinism, and defect classes (hallucinations, bias, prompt injection);
• designing evaluation rubrics and prompts, and the LLM-as-judge method;
• automated eval pipelines: promptfoo, DeepEval, pytest;
• testing RAG systems and AI agents (tool use, tracing);
• basic red teaming: prompt injection and jailbreak tests;
• the regulatory context: EU AI Act, ISO/IEC 25059.

Course highlights

• 120 academic hours, of which 64 are classroom contact hours, with practical work;
• in person in the centre of Tallinn (Tartu mnt 18) and/or online; training materials are included in the price;
• price 1700 € (VAT included);
• official certificate upon completion — with a “pass” grade on at least 75% of the homework assignments;
• no prior programming experience required: the course includes a Python foundations block.

Who this course is for

• current and aspiring testers who want an in-demand specialisation — testing AI systems;
• professionals from other fields considering a move into the AI industry through model quality assessment;
• linguists, editors, translators and analysts — working with language and attention to detail (rubrics, labelling, response evaluation);
• developers and engineers who need to verify the quality and safety of LLM features in their products.

Instructors: Roman Kutselepa and Nikolay Zubrilov.

Book a free consultation — we’ll walk you through the program, format and payment options.

If you know someone who’d find this relevant, share the course link with them.

When companies like Wise, Swedbank and Veriff hire people who can turn raw data into decisions, they look for exactly the skills this programme builds.

The Data Analysis with Python course opens a new group on 17 August 2026 — in person in Tallinn, with session recordings for online participants. No prior programming experience is required: you start from the basics and finish able to build real reports.

The programme covers Python (Pandas, NumPy), SQL and relational databases, data visualisation (Matplotlib, Seaborn, Plotly) and tools such as Jupyter, Google Colab and Streamlit.

It is taught by practising data professionals — Igor Yeryomenko (Senior Finance Data Analyst, Bolt) and Nikolai Zubrilov (Python/AI developer, Mentastic) — who work on live projects, so the cases are applied rather than theoretical.

Graduates have gone on to analyst and engineering roles at companies including Wise, Swedbank, Veriff, Betsson, Omniva, Inbank, MAXIMA and If Insurance.

Course funding options: details on the FAQ page.

If you know someone this might be a fit for — share the course link with them.

We keep an eye on the tools used across our courses. Two more updates from recent weeks.

• Figma (June 2026) — you can now bring a website onto the canvas as structured layers: the Chrome extension copies a full page or selected elements and pastes them into your design to study and rework. Handy on the UX/UI design course (details).
• Docker Desktop (June 2026) — updates to Docker Offload, Engine and Buildx, support for time namespacing in ECI protected containers, and improved stability on macOS and Windows. Containerisation is a core skill on the DevOps course and the cloud services course (details).

We keep our programmes current: as the tools update, so do our practical assignments — so the skills stay in demand with the market and employers.

We keep an eye on the tools we teach and keep our programmes current. Here are two notable updates from recent weeks — and how they relate to our courses.

• Claude Fable 5 — 9 June 2026. Anthropic released its new flagship model: always-on adaptive reasoning, a context window of up to 1 million tokens, and strong results in software engineering, data work and image analysis. For anyone learning AI tools on the Modern Neural Networks and Their Applications course, it sets a new benchmark for what large models can do and how to put them to work on real tasks.
• Figma AI Agent and Config 2026 announcements — 22–23 June 2026. At its Config conference, Figma introduced a built-in design agent, an expanded Figma Make (connecting to your codebase and shipping changes via pull request), and Design System Intelligence features for keeping design systems consistent. We cover these capabilities on the Prototyping in Figma and UX/UI Design course, so you work with current tooling hands-on rather than in theory.

Tools change fast — we update our programmes so the skills stay in demand with the market and employers, whether you are starting out in the profession or growing your career.

A short roundup of recent releases for tools used on our courses — we keep our programmes current.

What's new

• PostgreSQL 19 Beta 1 (4 June 2026) — the first beta of the new database branch, open for testing ahead of the final release in autumn 2026. PostgreSQL beta page.
• Postman (June 2026) — a new built-in Browser Tool lets you open a web app inside Postman, inspect page content and network traffic, and use Agent Mode to automate workflows and generate tests. Postman release notes.
• Figma (June 2026) — the Design agent gained web search for live context inside the file, and Runway's Aleph 2.0 video model arrived on the Weave canvas for frame-level edits. Figma release notes.

These tools are part of the hands-on work on our Data Analysis with SQL, QA: manual and automation and UX/UI design courses. We keep the material in step with each version — skills in demand with the market and employers, for a start in the profession or a step up within it.

Demand for cloud and DevOps skills keeps growing in 2026 — a market signal worth noting whether you are starting out in the field or growing within it.

According to CIO's “10 hottest IT skills for 2026”, AWS skills appeared in roughly 14% of job listings in 2025, up from about 12% a year earlier, while CI/CD requirements rose to around 9% of listings. Hiring intent is strong too: Robert Half research reports that 78% of technology leaders plan to increase permanent headcount in the second half of 2026.

Cloud and automation are becoming a baseline expectation in hiring rather than a niche add-on — in demand with the market and employers alike.

We keep our programmes aligned with that demand. Our DevOps, Linux and CI/CD course covers the core toolchain — for a start in the profession or a step up within it.

The main format is classroom training in Tallinn, with recorded sessions and an online option available.

According to OSKA, by 2027 Estonia's ICT professions should employ at least 47,600 people — demand for roughly 18,000 new specialists, around 2,600 every year. ICT remains the country's largest skills-shortage area and one of its best-paid fields, with an average salary above €3,400 per month.

The same study points to programming, data analysis and project management as the most sought-after areas. This is steady, structural demand rather than a short-term spike: skills in development, data, DevOps and cloud are valued by the market and employers alike — whether you are starting out in the profession or growing within it.

Generated by Routine — 2026-06-23 02:50 (Europe/Tallinn)

A short round-up of recent releases we track for our development and testing courses.

• GitHub Copilot App (2 June 2026) — the first standalone agent-native desktop outside the IDE: an overview of active sessions, tasks and pull requests across repositories. Source.
• GitHub Copilot SDK — general availability (2 June 2026): embed the Copilot engine into your own services, with tool calling and streaming sessions. Source.
• Postman Agent Mode (1 June 2026) — an AI agent for autonomous API testing workflows. Source.

We keep the course toolset current — part of preparation that is in demand with the market and employers.

Text generated: 23 June 2026, 03:26 (Tallinn).

In 2026, 38% of office workers use generative AI every day — up from 11% in 2024. Skills demand is following: the share of job postings requiring AI tool proficiency has grown from around 5% to approximately 9% in just one year.

This is a market signal, not a temporary spike. Applying AI in day-to-day work is becoming an expected skill — both for those entering a profession and those growing within their current one.

Generated by Routine — 2026-06-23 02:50 (Europe/Tallinn)