Automating Biochemical Experiments for a Silicon Valley Innovator

LeverX developed a comprehensive platform that fully automated the procedures for biochemical experiments, which allowed the client to more than double the number of experiments conducted within the first six months and shorten the synthesis cycles for new compounds.

Client and Challenge

The client is a Silicon Valley innovator, with a division specializing in discovering new methods for synthesizing biochemical compounds. Their key customers are pharmaceutical companies and defense-related organizations, which places high demands on the speed and reliability of results.

Before contacting LeverX, the client conducted biochemical experiments manually, which, as their business grew, became increasingly time-consuming and costly, creating a number of critical problems. The client was in a critical situation, having only 12 months to launch the solution under a government grant and with a sufficiently limited budget.

Key challenges:

• Slowdown in Target Molecule Development: Manual synthesis workflows, which required numerous iterations and complex calculations, led to long development cycles for new compounds.
• High Labor-Intensity and Limited Scalability: The manual nature of experiment design and execution limited the scalability of the scientists' work and created bottlenecks in the discovery processes.
• Reproducibility and Data Quality Issues: Manually designed and executed reactions were often not reproducible, especially in different labs with different equipment, which affected the reliability and accuracy of results.
• Lack of a Single Database: Data obtained during experiments was not systematized in a single database, which prevented using previous findings to accelerate new research.
• High Risk of Human Error: The complexity of multi-step synthesis made the process prone to human errors at any stage, which could lead to significant delays or require a complete restart.

Methodology and Approach

Faced with these critical problems, the client turned to LeverX. To respond promptly to the critical situation, the client organized a two-day workshop where our team presented a high-level architecture for the future solution. The client was impressed with our meticulous approach and clarity, and therefore entrusted us with the further development of the project.

The main task for us was to develop and launch the solution within a tight deadline and budget. To make this possible, LeverX applied an Agile approach, dividing the development process into the following five phases: 

1. Discovery and Requirement Gathering: We actively engaged with the client to clarify goals, expectations, and success metrics, performed in-depth research and feasibility studies, and conducted workshops to gather requirements and refine the project scope.
2. Solution Design: We leveraged our cross-functional expertise, including BAs, Solution Architects, and UI/UX designers, to develop wireframes, user journeys, and interactive prototypes for early validation. Next, we created scalable system architectures with a focus on future enhancements.
3. Agile Development & Delivery: We organized multiple Scrum teams with distinct responsibilities, such as UI/UX, business logic development, and system integration, delivering functional modules in sprints with continuous feedback from the client.
4. Quality Assurance & Testing: Our team performed automated and manual tests, including unit, integration, and performance ones. We validated data in real-time, which involved processing and verifying large telemetry and mass-spectrometry data sets.
5. Deployment & Support: At this stage, the implementation of the solution and its further support were carried out.

We were able to achieve the client's goals, including:

• Automating biochemical synthesis procedures from start to finish to reduce the time required for molecule design, setup, execution, and analysis of experiments.
• Optimizing the use of robotic hardware, allowing for up to five experiments to be conducted simultaneously, which increased productivity.
• Creating a platform capable of handling increasing workloads and complex biochemical synthesis processes.
• Seamlessly integrating with existing systems, such as formula design tools and mass spectrometry software, to create a unified workflow.
• Developing a platform that could be deployed and tested completely remotely, which solved the problems caused by the COVID-19 pandemic.
• Successfully implementing the solution within the allocated budget and tight deadlines.

Solution

Based on the architecture presented during a two-day workshop, and impressed with our approach, the client entrusted us with the project’s implementation. Under a tight deadline and with a limited budget, LeverX developed and delivered a market-ready solution—a platform that fully automates the biochemical experiment cycle.

Key solution modules:

• AI-Powered Route Planning: AI algorithms suggest optimal synthetic routes, reducing the time and resources needed for design.
• Experiment Automation: The platform translates JSON configurations into commands for robots, integrating with proprietary robotic complexes.
• Parallel Execution: The system allows running up to five experiments simultaneously, which significantly maximizes hardware utilization and throughput.
• Mass Spectrometry Integration: Analysis results are automatically imported into the platform, reducing manual data handling.
• Telemetry Data Collection and Processing: The platform collects and processes telemetry data from sensors in real-time, providing effective visualization and analysis.
• Automated Reporting and Feedback Loops: Processed data is converted into detailed reports. Based on insights, the system can suggest parameter adjustments or new experimental configurations.
• Unified Database: The creation of a single database allowed for the systematization of experimental data, which provided valuable analytical insights that were previously unavailable.

Technology Stack

The platform was developed using the following technology stack:

• Database: PostgreSQL.
• Backend: Python (for core business logic) and GoLang (for agents).
• Frontend: React / Redux.
  

Results

LeverX provided the client with a well-thought-out solution. The project was successfully launched within the agreed-upon timeframe and budget.

The client achieved the following measurable results:

• Increased Throughput: Within the first six months after the solution’s launch, the number of completed experiments more than doubled due to workflow automation and parallel task execution.
• Accelerated Time-to-Discovery: The automation of design, execution, and data processing significantly reduced manual work and sped up synthesis cycles.
• Enhanced Accuracy and Reproducibility: Real-time data collection and automated reporting minimized human error and increased the reliability of results.
• Creation of a Single Source of Truth: Thanks to data centralization, analytical insights that were previously unavailable became the foundation for decision-making.
• Overcoming Constraints: LeverX successfully met the challenge of developing and launching the solution under a very tight deadline and with a limited budget.