Research | Oliveira Lab

Research Interests & Support

• AI-enabled synthetic microbial communities
• Long-term live-cell imaging and spatial-temporal dynamics
• Automated microfluidics and high-throughput discovery

• Machine learning, computer vision, and predictive modeling
• Digital twins and closed-loop biological control
• Intelligent biosensors, organ-on-a-chip, and biomanufacturing

AI-enabled discovery and engineering of microbial communities

Microbial communities exhibit complex behaviors across space, time, and environmental conditions. We combine synthetic biology, long-term live-cell imaging, automated microfluidics, and machine learning to quantify these dynamics and identify design principles for programmable community behavior. Controlled, spatially resolved datasets support predictive models of gene expression, intercellular communication, growth, and collective function.

Rather than treating AI only as a downstream analysis tool, we are building closed-loop experimental systems in which computer vision, predictive modeling, and digital twins help select the next experiment, recommend perturbations, and guide living systems toward desired states.

Closed-loop microenvironments, living devices, and bioreactor networks

We co-design cells with the microenvironments in which they operate. Our platforms integrate microfluidic screening, biosensing, automation, and control to study long-term behavior and create intelligent living devices for health, agriculture, environmental monitoring, and biomanufacturing. Emerging directions include autonomous discovery platforms, organ-on-a-chip and tissue-engineered systems, and AI-guided bioreactor networks connecting real-time measurements, models, and interventions.

Support

Engineering on-chip very-large programmable microbial communities with complex traits

Microfluidic-based screening platforms help find and characterize cell candidates with desired functions. For example, a library of cell-based biosensors that sense concentrations of heavy metals can be determined for a micro-environment while external variables are tested (e.g., inorganic metals and metal alloys). Depending on the complexity of the application, though, a community of cells may be required to perform a complex task. Other microfluidic devices that can sustain cell growth for long-term measurements are required in these cases, and many devices are available in academia and industry. Our lab investigates how microfluidic primitive properties (i.e., structure and dynamics) can be scalable to larger networks of primitives in a predictive way when implementing monolayer and multi-layer (with complex routing and microfluidic technology integration) fabrication approaches.