From billions of parameters to millions of documents, here’s how we’re building the future of property auction analytics—powered by HPC, LLMs, and predictive modeling.
Published on: April 24, 2025
Author: Renato Mutavdzic, Jr. AI Engineer
Context: The Real Estate Auction Problem
Real estate auctions are an opaque and fragmented landscape. Legal documents are dense, auction listings vary wildly in format, and market predictions are speculative at best. Investors and institutions often navigate this chaos manually.
BidSale, our AI-native platform at VendueTech, aims to solve this through automated document analysis, real-time price forecasting, and retrieval-augmented recommendation systems—powered by LLMs, deep learning, and massively parallel computing.
HPC: Not Just Nice-to-Have, But Mission-Critical
Training models like LLaMA 2, Falcon, and even domain-specific LLMs for legal/real estate use requires an environment beyond cloud or local resources.
Enter EuroHPC’s MareNostrum AI Cluster, where we will run:
Monthly usage: approximately 2500 node hours
Total dataset for training: approximately 2 TiB
We’ll leverage this to:
The AI Stack: Modular, Multi-Modal, Multilingual
Natural Language Processing (25%)
Predictive Analytics (20%)
Machine Learning (20%)
Computer Vision (25%)
Data Mining & Pattern Recognition (10%)
What Makes BidSale Unique
FER TakeLab: Academic Precision at Production Scale
We’re not alone in this journey. Our academic partner TakeLab at the University of Zagreb’s Faculty of Electrical Engineering and Computing (FER) brings decades of research excellence in NLP, knowledge grounding, and language modeling.
Their role:
This industry–academia synergy helps us scale fast without sacrificing rigor.
SRCE and the Role of National Infrastructure
In addition to EuroHPC, BidSale is supported by SRCE (University of Zagreb University Computing Centre), Croatia’s leading institution for national academic HPC infrastructure. SRCE is providing foundational support in terms of environment setup, containerized deployment compatibility, and compliance with national infrastructure requirements.
Moreover, we’ve applied for a CROBOHUB++ Innovation Voucher to further accelerate our work. This co-financing mechanism is intended to support high-potential AI innovations in Croatia and the EU, and we’re currently awaiting final determination.
This effort is also a precursor to our broader vision: joining an EU's AI Factory, a federated, production-grade, domain-specific platform for applied AI solutions in proptech, legal tech, and financial analytics. BidSale is only the beginning.
Performance Engineering & I/O Strategy
To mitigate common AI bottlenecks:
The I/O architecture allows data-hungry training loops to fully utilize compute without latency gaps.
Ethical AI: Fair, Transparent, Sustainable
Our Hypotheses (and Why They Matter)
We believe:
If confirmed, this transforms real estate—from bureaucratic chaos to real-time, data-driven decision making.
What’s Next
We’re not just building a product. We’re building the infrastructure for the next decade of real estate.
Read more at https://rnca.fccn.pt/en/marenostrum-5/
If you are an AI Engineer, PhD researcher, Algorithm developer, Quant, Back-end, Front-end engineer, or another field, check out our open positions and apply here.
Subtitle: Early assessment highlights excellence, feasibility, and impactful potential of VendueTech's AI-powered distressed property auction platform
Published on: April 22, 2025
Author: Ivan Livić, CEO & Co-Founder
VendueTech is proud to share that our project, Bidsale: A Next-Generation AI-Powered SaaS Platform for Distressed Property Auctions, has received strong preliminary results in the Eureka funding assessment.
Bidsale is a SaaS solution designed to transform how distressed properties are evaluated and auctioned across Europe. By combining automated valuation, predictive legal risk analytics, and real-time market data, Bidsale empowers banks, investors, and legal entities to streamline judicial property sales.
The Eureka panel provided detailed feedback across 11 dimensions. Across three independent assessments, Bidsale received scores of 18/18, 15/18, and 16/18, giving us an impressive:
These results reflect the consistency of Bidsale's excellence and the trust in its technical and commercial feasibility.
The panel commended the consortium of three partners, noting their complementary strengths in software development, system integration, and project experience. While the panel recognized the technical ambition of the platform, they acknowledged the team's strong ability to execute.
The assessors highlighted the realism of our market projections, especially our focus on judicial auction inefficiencies. Bidsale's ability to reduce transaction time, increase investor confidence, and improve recovery rates for banks was seen as highly relevant.
The project plan was described as realistic, well-structured, and credible, with clear goals, logical work packages, and milestones. Our partners’ prior experience with both national and international R&D projects was seen as a strong asset.
The panel praised Bidsale's novel approach to applying AI in property valuation and cross-border legal harmonization. The use of predictive models for legal and financial risk was noted as ambitious and forward-thinking.
The main challenge lies in integrating legal frameworks across EU countries. While technically feasible, the complexity of legal harmonization remains a major variable to monitor.
This assessment confirms that Bidsale is not only innovative, but also grounded in a well-planned and feasible execution roadmap. With clear commercial and societal value, we are one step closer to enabling faster, safer, and more transparent distressed property auctions.
We're grateful to our partners and advisory teams for helping shape a product with real European impact.
Next steps: We are currently preparing for the final phase of grant evaluation. Meanwhile, we continue to build partnerships, expand our data coverage, and iterate on our product with early adopters across the EU.
Stay tuned for further updates from the frontlines of real estate innovation.
Given that we love a good challenge and creating tech solutions that improve the quality of life and address real issues, we participated in Hackl - the Hackathon for Open Zagreb at the end of November.
The theme of the Hackathon was the reservation and overview of available city spaces in Zagreb.
Zagreb has 225 public spaces across 17 neighborhoods that aren’t optimally utilized. A significant number of citizens aren’t aware they can book city space, the ones they are facing significant challenges trying to do so.
The current presentation of space information is poor. It does not help in finding spaces and available time slots, the process of booking is manual and inefficient.
Therefore, Ivan Mikuličić , Ivan Livić, Kresimir Certic Misch, and I attempted to solve this problem within the given timeframe of about 40 hours.
2. Admin interface (with several different levels of permissions) - To enable City employees to monitor, approve, reject requests, and add new spaces with relevant details
We started by:
• Briefly analyzing and discussing the functionalities specified in the application requirements (there were indeed many – precisely 12!)
• Brainstorming initial ideas and challenges we foresee based on our experience
• Assigning responsibilities based on personal skills
Given the very limited timeframe, coding the solution "from scratch" was unrealistic.
To find a potentially workable solution, we decided to explore no-code and low-code platforms. Ivan Mikuličić and Ivan Livić analyzed several platforms, including Coda, Notion, Airtable, and Bubble.
After a quick evaluation, we decided to use Airtable - due to its user-friendly interface, logic, and functionalities it supports. It seemed like a platform that supported everything we needed to implement.
Advantages of Airtable:
Since we couldn't figure out how to brand the Airtable application and create a custom UI in a short timeframe, I proposed to design another solution in Figma.
This solution was a vision for what we could do in the future - if we were given the chance to execute this project.
I decided to start with the experience from the Citizen’s point of view. Assuming that they would be searching for spaces and submitting requests 'on the go,' I focused on a mobile web solution.
Custom design is powerful because it allows us to:
A few hours into the project we had an application name, logo, colors, and icons in the making. It was slowly coming to life!
The clock was ticking, we were all busy working, but somehow the progress seemed slow. We were in the problem-solving, solution-building phase, going somewhere but very unsure if we would reach our destination.
I started feeling tired, losing focus, and not sure of my priorities. Not having the time to play with potential solutions and explore different directions before setting to one was frustrating.
We did a first round of internal review and feedback. I have to say - I didn't like what I heard 🙂 I generally handle the feedback fine as a designer. But this time it frustrated me. I knew the output wasn’t great, I had the vision for what I wanted to do - but the time was so short. I felt a bit defeated.
Once Ivan Mikuličić was clear on the requirements he isolated himself as much as possible from the noise and focused on execution. (good call! :D)
He managed to achieve what I thought was impossible – in 12 hours he developed all the required functionalities for the user and admin interfaces on Airtable.
We tested the solution several times, and with minimal changes to colors and component positions, it looked really good!
It's an understatement to say that we were amazed by the possibilities and usability of Airtable.
I designed the mobile responsive web application for the Citizens in Figma and linked it to a clickable prototype.
Video of the Figma design
No, we didn't win. But we are genuinely proud of the solution we managed to create during the Hackathon! Until this Hackathon, we haven't fully explored the possibilities of low-code and no-code applications in creating digital products.
We are now convinced that they are indeed a good potential alternative to coding from scratch (for certain purposes of course).
🗣 🥇 Happy to announce that VendueTech innovative product idea not only was accepted, but came 1st in the ZICER - Zagreb Innovation Centre's new cohort of startups.
🎉 Thanks to awesome team for getting another positive milestone!
If you are interested in a equity based role in early stage startup, we are currently hiring for:
AI engineers
Full-stack engineers
Front-end developers
Back-end developers
UX Researcher
Sr. Product designer
Legal
Investor Relations Manager
Reach out to jobs@venduetech.com with a CV and a short motivational paragraph.
At ZICER - Zagreb Innovation Centre, a new generation of innovative technology startups presented interesting solutions that, with the application of achievements in the field of robotics and artificial intelligence 🤖 , have the potential to solve the challenges of cyber security, health, energy, education, and a wide range of other areas.
VendueTech's CEO Ivan Livić presented our startup and we are very excited to be part of this year's competition. 💪
Startup Factory will provide technology-oriented teams and young startups with accelerated development and help with validation, entering the market, and accessing the first customers and investors. All selected participants will gain valuable knowledge through an intensive schedule of boot camps consisting of workshops, lectures, and mentoring. The program ends on December 19, 2023 with a demo day at the startup conference Zagreb Connect. Then, the expert jury will select the three best teams or startups from all the participants, which will win funds from the money fund of 30,000 euros. 💶
