Misha

Serial entrepreneur and founder of Author.com, an AI tool that streamlines academic writing.

Meet Misha: Twenty Years from Code to Four Start-Ups

Please tell us a bit about yourself. What experience do you have and what are you working on now?

I was born in Astrakhan and now live in Luxembourg. My background is quite diverse: twenty years ago I started out as a programmer and did well in that field. In 2008 my friends and I began building websites on demand, and by 2012 I launched my first start-up. To date I have founded four start-ups, achieved one exit, and completed three funding rounds.

What inspired you to create Author.com? Was there a specific moment that sparked the idea?

When ChatGPT 3.5 was released, the team at SKL.VC—where I was a managing partner—started thinking about how we could use the new technology to build something fresh. That launch was the trigger that pushed us toward Author.com.
From first code to four startups

Could you tell us more about your move to Luxembourg—when did it happen, and what were your main reasons and goals?

We moved here a year and a half ago after living in Israel. The primary reason was to find a new home that felt safer for our children.

How has relocating to Luxembourg affected Author.com and your personal life? What new opportunities or challenges has it brought?

For Author.com nothing really changed because our team has always been fully remote. Personally, though, life improved a lot—we can make plans again and feel far less stress. Luxembourg is a small country with big ambitions around start-ups, so networking and business opportunities have actually grown.
The Spark Behind Author.com: Riding the ChatGPT 3.5 Wave

Which of your ventures would you like to talk about today, and what is its name?

I’d like to talk about Author.com.

Could you describe the specific moment of inspiration that pushed you toward creating Author.com after the release of ChatGPT 3.5? What exactly lit the spark?

When ChatGPT 3.5 came out, I was still a managing partner at SKL.VC. We gathered in the office and asked ourselves how we could turn this new model into a business. It was obvious the technology would transform many industries, so we mapped out which ones would change first and where the model was already good enough. Very quickly we realised that text-heavy tasks—especially academic writing—were ripe for disruption. That insight became the starting point for Author.com.
Text is the first thing to change

What was the toughest challenge you faced in the first months of launching Author.com, and how did you cope with it?

At that time the technology simply wasn’t ready for serious research work. The models hallucinated far too often, which was a deal-breaker for academics, and the context windows were tiny, so we couldn’t feed in all the source material. To move forward we bought a dataset of eight million research papers and started building retrieval tools around it.

What proprietary solutions did you develop to overcome hallucinations and the limited context windows? Could you walk us through the process and the results?

We built our own Retrieval-Augmented Generation (RAG) model that pulls information only from the research papers we upload. To get around the context-window limit, we generate the output in small, consecutive chunks and stitch them together seamlessly; our technology keeps the narrative coherent across those blocks.
Taming Hallucinations: Inside Author.com’s RAG Architecture

Which in-house solutions did you have to build to stop AI hallucinations and work around the tiny context windows? Please walk us through what you did and what you achieved.

Technology simply wasn’t ready for writing research papers: the LLM hallucinated a lot, which was a complete deal-breaker, and context windows were still extremely small. We bought a database of eight million peer-reviewed papers and built our own Retrieval-Augmented Generation pipeline that pulls facts only from this trusted corpus. To get past the narrow output window, we generate the text in small segments; our own middleware stitches those segments together seamlessly while keeping the narrative coherent.

From your perspective, what is happening in AI-driven scientific research today that most people still don’t notice?

The changes unleashed by AI tools are much bigger than they look. I have already replaced analysts and a good part of my developer team, and I still think the technology hasn’t reached its peak. Very soon the world will change—both business and everyday life. Behavioural models are shifting, some professions are becoming obsolete, and before long we’ll see robots entering our daily routines at full speed.
“Hallucinations were a total deal-breaker.”

Could you explain in simple terms what Author.com does and why it matters?

Our AI assistant helps students and researchers produce scientific papers much faster. By grounding every answer in verified literature, we let academics focus on ideas instead of formatting and citation drudgery.

What would you say to someone who has dreamed of starting their own venture but still hesitates to take the first step?

Just try. Water wears away stone; if you dedicate even half an hour a day to your project, you will inevitably see results.
Relocating to Luxembourg: Safety, Stress Relief and Startup Perks

Misha, could you share a significant recent event in either your business or personal life that you think is worth highlighting?

The most important thing that happened recently was our family’s relocation from Israel to Luxembourg.

Tell us a little more about that move—when did it take place and what motivated the decision?

It happened about eighteen months ago. We were looking for a new home primarily to ensure our children’s safety. Luxembourg felt right: politically calm, very international and only a three-hour flight from almost anywhere in Europe.
“Less stress, bigger dreams.”

How has living in Luxembourg changed things for Author.com and for you personally?

For Author.com, not much changed at all—our team has been remote from day one, so operations continued smoothly. For my personal life, however, the impact was huge: we could finally make long-term plans and the everyday stress level dropped dramatically.

What do you see as the main advantages of Luxembourg’s startup ecosystem?

We suddenly found ourselves in a small country with very big ambitions. The government actively bets on startups, so networking is easier, and there are plenty of programs and events that open doors for both funding and partnerships.
AI’s Near Future, Simple Startup Advice and a Rapid-Fire Glimpse

Looking ahead, how do you see your industry changing, and which AI-driven trends are still flying under the radar?

The changes brought about by new AI tools are far bigger than most people realise. I have already let go of analysts and a large group of developers, and I still believe the technology hasn’t peaked yet. Very soon the world—both business and everyday life—will look different: behaviour patterns are shifting, certain professions are becoming unnecessary, and in a little while we’ll watch robots enter our lives at speed.

What is your favourite morning drink?

Orange juice.
“Just try for 30 minutes a day.”

Which living person would you most like to have coffee with?

Elon Musk.

What is the best purchase you have made for under $100?

An orthopedic pillow—it really makes a difference.