Building Bioloupe's Data Brain with AldAstra

Imagine slashing biopharma deal-screening time from months to mere weeks. That's the impact AldAstra delivered for Bioloupe, with our work estimated to cut this critical process by as much as 60%. The biopharma sector is a maze of information. Finding the right drug, trial, or partner means sifting through mountains of data. Bioloupe aims to bring clarity and speed to this complex world, and we at AldAstra partnered with them to build the core intelligence engine.

The Partnership

Our collaboration with Bioloupe was a shared mission.

Bioloupe was tackling a fundamental, high-stakes problem: making biopharma deal-making more efficient. "They weren't looking for just another dashboard," recalls Synim, our co-founder and lead data architect on the project. "They wanted to fundamentally change how quickly researchers could connect disparate information to find real opportunities. That's the kind of challenge that gets our team genuinely excited." This focus on meaningful impact resonated deeply with us.

The technical task itself was compelling: transforming oceans of unstructured data (publications, trials, patents, pipelines , regulatory policies) into a resource as searchable as Google and as reliable as PubMed. We thrive on complex data puzzles, and solving them sharpens our own AI infrastructure and large-scale data processing capabilities.

Beyond the tech, the human element was key. The Bioloupe founders brought priceless pharma experience; our team brings a passion for building game-changing products. This union created the perfect environment for innovation.

Engineering the Data Intelligence Engine

Our team built the heart of the Bioloupe application – a sophisticated data foundation and AI-powered intelligence layer, creating a cohesive, intelligent system that could anticipate user needs.

1. Data Pipelining

We developed extensive pipelines to gather information from over 40 distinct sources. This system actively cleans, structures, normalizes, and links this disparate information daily. Using advanced Natural Language Processing (NLP) to understand the text and custom graph analytics to map connections, it reveals insights previously hidden within the data silos. The entire architecture was designed for scalability, ready for exponential data growth. Imagine trying to piece together a million-piece puzzle with pieces from 40 different boxes – our systems automate that, and then show you the hidden picture.

2. Powerful Semantic Extraction

On top of this foundation, we implemented powerful semantic extraction tools. This includes a Dual-encoder structuring model, which essentially learns to understand the meaning behind complex biomedical questions, not just keywords. So, a question like "Find Phase II IL-2 agonists for solid tumours" gets precise, relevant insights in milliseconds. Users can also input identifiers like company tickers or clinical trial IDs, and the engine uses semantic extraction across the vast integrated dataset to distill concise, high-value insights that surface potential competitors, mechanistically similar assets and relevant biological pathways.

3. Actionable Intelligence & Sustainability Ops

To ensure the insights were immediately usable and the system sustainable, we focused on:

• •Intelligent Opportunity Ranking: Algorithms that sift through complex signals to highlight the most promising leads.
• •Ops-Friendly Dashboards: Clear, intuitive interfaces for data evaluation, rapid evaluation and QA decisions.
• •Automated MLOps: We implemented what we called “The Data Consistency Layer” with automated workflows for data refreshes and flexible model updates.
• •Comprehensive Handover: We worked side-by-side with the team at Sogody, who masterfully took the reins of the backend, cloud infrastructure, and front-end development. Our detailed runbooks and pair-programming sessions ensured they were fully equipped to operate and extend the system.

Impact and Growth for AldAstra

Our work with Bioloupe was a landmark for us. It deepened our expertise in the life sciences vertical, and the project's success has already sparked interest from other pharma companies. The technical innovations, particularly in our dual-encoder models and graph-based ranking, are now enhancing other AldAstra products.

There's also considerable team satisfaction. As Diellëza, our co-founder who also worked hands-on with the Bioloupe project, put it, "It's incredibly rewarding to know that the systems we build can directly contribute to researchers' efforts in discovering the next breakthrough drug. That's what makes this work so meaningful for us." And yes, perhaps led to a few more biotech memes internally.

Project Path and Ongoing Partnership

From initial research in early 2024 through phased development milestones, we completed the full system build. The strategic handover of the operational system, including backend and frontend components, to the Sogody team was finalized by Q4 of 2024. Today (April 2025), AldAstra remains in close contact with all teams and continues to offer support wherever needed, driven by goodwill and our strong ongoing relationship.

A Partner's Perspective

What's Next for AldAstra?

The Bioloupe project sharpened our tools and reinforced our passion for solving complex data challenges. We're now channeling that same expertise, rigor, and collaborative spirit into our own suite of products. We're applying the lessons learned and the high standards achieved to build solutions designed to make a significant impact.

Our latest product, PayStorm, just launched—built with the same obsession for structure, clarity, and speed. It’s our solution to the chaos of accounts receivable — a bold stand against late payments, once and for all. If you're curious what that looks like in practice, take a look. Same focus on quality, same drive for innovation. Slightly bigger dreams.