I work across product, marketing and business leadership.

Early interests and Shipitwise

I began blogging and exploring ecommerce in 2007. Writing helped me work through ideas, while my interest in technology came partly from my great-grandfather, an independent inventor. A good physics teacher also made science interesting enough that I considered studying it at university. I have kept that curiosity throughout my working life.

In 2015, I attended Draper University in San Mateo and became involved in Shipitwise. I helped the company get into Startup Wise Guys and took on marketing. My father and the other founders knew the shipping problem firsthand, and we served our first thirty customers through a website form before we had a software product.

We later built tools that combined carriers into a shipping chain. I worked on marketing, partnerships and crowdfunding, while the engineers built the pricing logic. Larger customer payments did not necessarily improve our margins or lead to repeat business. After we halted operations in 2019, I returned to run the exit process into the Finnish company Speys, then continued with marketing and investor relations. The technology had value, but we needed another business to take it forward. That experience made me look more carefully at the economics behind a promising product.

Robotex and international partnerships

I joined Robotex as Chief International Officer and later became Chairman. It was already an established robotics festival, built by universities, students and many others. I worked on international promotion and partnerships, helping grow the franchise and robotics-education network to 15+ countries, including China, Colombia and India. Much of that growth depended on local organisers who understood their schools, communities and partners.

In Tallinn, we brought competitions, workshops, technology companies and a conference together. I enjoyed being around people building hardware and giving families a chance to try it. The festival could introduce a child to robotics while giving people in the industry a place to discuss their work.

I also pushed too hard to turn Robotex into a startup-development organisation. Its educational role was already well understood, and my broader ambition began to confuse that. We chose to focus on education, and I stepped away from operations. I learned to give more weight to the purpose and strengths of an existing organisation before trying to expand its role.

Springrise and hands-on marketing leadership

Lele and I ran Springrise from 2019 to 2022. As managing partner, interim CMO and team lead, I worked with more than 20 companies across manufacturing, technology, finance and other industries. The work involved business models and marketing strategy as well as writing, websites, digital channels and ongoing management. We continued while WOF Labs was taking off, then wound down the agency to focus on it.

With Tech Group, we clarified machine building and factory automation as distinct services. With Krakul, we worked on IoT and autonomous-technology positioning, technical articles and customer stories. Punk Finance, Monestro and Cachet brought lending, investment and insurance into the work; Flowit and Worksup added experience in software and business services. I enjoyed understanding the technical detail well enough to explain the problem, alternatives and responsibilities to a buyer.

I have done marketing myself, built teams and worked with outside specialists. The right mix depends on the business and the people available, and a partner's favourite channel may not be the one customers use. Missing input and slow approvals often cost us time, so I became more deliberate about agreeing who was responsible for each part of the work. Taking an interim leadership role still meant writing, analysing and working through the details with the team.

Kiri and semantic search

In spring 2020, before GPT-3 was announced, I started exploring semantic search. I was frustrated by elaborate folders and duplicate knowledge bases: I wanted to find information without knowing the exact keyword someone else had used. At Kiri, I led research, sales and operations alongside four engineering co-founders. We built an MVP, tested it with a small beta group and gave demonstrations to the Estonian Government and Tele2. We did not reach customer deployment.

Interviews with consultants, finance professionals and lawyers showed how much access rights and security mattered. Some buyers could also get adequate results much more cheaply elsewhere. We struggled to differentiate, and the team's preferred machine-learning challenge did not quite match the search problem customers needed solved. We eventually wound the venture down.

Kiri shaped how I assess applied AI. I want to understand which task it improves, what information it needs and who checks the result. Often, AI makes sense as a feature within a product people already need. A useful demo gives us something to test, but we still have to establish whether customers will pay and whether it can work reliably in everyday use.

WOF Labs, Supplain and BananaConf

At WOF Labs, DeShip developed into Supplain, a way for companies to exchange supply-chain information without replacing their existing systems. World of Freight brought together a community and NFT sales to finance development. Our six-month sales campaign brought in over €1M for that work. This was sales-based financing, rather than an equity investment round. We delivered games, research, architecture and initial infrastructure, but did not complete Supplain.

I let games, a protocol and events compete for the same people and money. Although the ideas were related, each needed its own owners and enough resources to finish the work. We improved our focus too late. NFT Tallinn later continued separately as BananaConf, where I worked on speakers, sponsorship and partnerships. We also hosted an innovation event in Milan with speakers from D&G, Prada and Valentino.

The events ended after May 2025. I value the relationships we built, though attendance alone did not make the business sustainable. Some established brands asked what they could actually use and returned to us when technology providers changed direction. Those conversations helped shape Perxify and gave me a better understanding of what remained unresolved after an event.

Perxify and running our own marketplace

Alvar Laigna and I began developing Perxify at the beginning of 2024, initially around loyalty programmes. Customer requests showed how much depended on the catalogue, accounts, orders and delivery around those programmes. I led business and product direction; Alvar led architecture and engineering. We built API-first commerce infrastructure and began developing our own book marketplace, Raamatuhai, also known as BooksShark, in spring 2025.

By December 2025, Raamatuhai was running on our own infrastructure. We used AI for catalogue enrichment, deduplication and exception handling, alongside Stripe Connected Accounts and DPD in live commerce. The marketplace included barcode- and photo-assisted listings, price suggestions, grouped offers and checkout that combined one seller's books into one shipment. We also had to handle payments, payouts, returns and buyer assistance. AI could identify a title, but sellers still needed to confirm the particular copy.

We first approached investors in September 2024 and were asked to return once we had two clients. We worked towards that, but a major prospective client had to cancel in spring because of limited funds. We then started building Raamatuhai. After its launch, we resumed fundraising in January 2026 and spent eight months on it without securing external funding. Of three potential funding leads, one cancelled, one lost funds in bad investments and one had to postpone until 2027.

We had been bootstrapping since 2024 and could not commit to customers without external financing for at least six months of runway. In October 2026, we halted active development and began closing BooksShark. Perxify remains a company. Our January sales offer covered complete stores, marketplaces and loyalty programmes; in May we shifted to API sales. By late August, we had agreed on Codex, Stock, Insights and Hugo plus a manufacturer storefront, but we did not complete that rollout.

I was too patient with financing conversations that kept progressing without money arriving. I still see value in Agentic Commerce, where AI helps people discover, compare and buy products. Running the marketplace gave me a much closer understanding of the product information, prices, availability and support that both people and their assistants need. Those details require care regardless of how a buyer reaches the checkout.

Law, regulation and physical products

I studied law at the University of Tartu in 2013–2016. Before that, debating, teaching debating and working on education policy helped me learn to research difficult questions and explain my reasoning. As an Estonian Web3 Chamber board member, I have worked with MiCA and the AI Act, and my wider regulatory research includes GDPR and PPWR. I use that understanding in product and business decisions, while formal legal opinions remain the responsibility of qualified specialists.

I am particularly interested in premium and luxury physical products, especially those that combine thoughtful design with robotics, IoT or AI. I care about materials, proportions and everyday use as well as connectivity and commercial viability. My role is in product and business development rather than electronics design, and I also work with companies whose products are entirely digital.

Across these projects, I have often worked with people whose expertise is different from mine. Understanding enough of their work to ask useful questions, weigh the trade-offs and take responsibility for a decision has been a large part of my job.