HPC/Cloud Infrastructure
Life sciences organizations running computational chemistry, genomics, AI, or large-scale analytics need infrastructure that was designed for science — not adapted from generic enterprise IT. BioTeam builds, optimizes, and modernizes HPC and cloud infrastructure for research organizations that can’t afford downtime, cost overruns, or environments that researchers won’t actually use.
Infrastructure-as-Code for HPC
Ad hoc HPC deployments don’t scale across organizations and don’t survive staff turnover. BioTeam builds fully automated, repeatable HPC environments using infrastructure-as-code so every deployment is consistent, documented, and reproducible.
How BioTeam does this:
- Full HPC environment defined in code — compute, networking, storage, scheduler, software stack
- Repeatable across multiple organizations or AWS accounts without manual configuration
- Version-controlled so changes are tracked and rollbacks are possible
- Dramatically reduces time-to-cluster for new environments or disaster recovery
Seen in practice: At a pharmaceutical organization, BioTeam deployed a complete enterprise HPC environment on AWS using IaC — collapsing what would have been weeks of manual setup into hours. Read the case study
AWS Landing Zone for Life Sciences
Standing up cloud HPC without a proper account structure creates security debt that compounds over time. BioTeam implements AWS Landing Zone architectures purpose-built for life sciences organizations — multi-account, compliant, and secure from day one.
How BioTeam does this:
- Multi-account AWS architecture using Landing Zone Accelerator with life sciences guardrails
- Service Control Policies (SCPs) that enforce security boundaries across accounts
- Fortinet NGFW integration for network security at the perimeter
- Compliance controls baked into the foundation, not retrofitted later
Seen in practice: For a pharmaceutical HPC deployment, BioTeam stood up a fully governed AWS environment with NGFW, SCP guardrails, and multi-account structure before a single HPC workload ran — ensuring the environment was audit-ready from the start. Read the case study
FlexLM + Slurm Integration
Computational chemistry licenses are expensive. When job schedulers aren’t aware of available license tokens, researchers hit failures mid-run, licenses sit idle while jobs queue, and compute costs accumulate with nothing to show for it. BioTeam wires license awareness directly into Slurm so the scheduler knows what’s available before it launches a job.
How BioTeam does this:
- FlexLM license server integrated with Slurm as a trackable resource
- Jobs only launch when both compute and license tokens are confirmed available
- Eliminates mid-run failures caused by license contention
- Reduces idle compute spend from jobs waiting on licenses that were never going to be available
Seen in practice: For a pharmaceutical organization running Schrödinger workloads on cloud HPC, BioTeam integrated FlexLM with Slurm to eliminate license contention — turning a chronic source of failed jobs and wasted spend into a solved problem. Read the case study
Cresset Engine Broker on Cloud HPC
Cresset’s computational chemistry tools weren’t designed for cloud HPC schedulers. There’s no existing playbook for running Cresset Engine Broker inside a Slurm-managed AWS environment. BioTeam built the integration from scratch.
How BioTeam does this:
- Custom integration of Cresset Engine Broker into a Slurm-based AWS ParallelCluster environment
- Job submission workflows adapted to Cresset’s licensing and execution model
- Tested and validated against real research workloads before handoff
- Documented so internal teams can operate and extend it independently
Seen in practice: For a pharmaceutical research team that needed Cresset running on cloud HPC alongside Schrödinger, BioTeam designed and implemented the integration with no existing reference architecture to work from. Read the case study
On-Demand GPU Notebooks
GPU and high-memory instances are expensive to run continuously. Researchers using Posit Workbench for interactive analysis don’t need those resources sitting idle between sessions — they need them available instantly when work starts and gone when it stops.
How BioTeam does this:
- Posit Workbench configured to provision GPU or high-memory nodes on session start
- Nodes terminate automatically when the session ends — no manual teardown required
- Researchers get the compute they need without IT involvement for every job
- Dramatically reduces GPU spend compared to always-on instances
Seen in practice: For a pharmaceutical organization running interactive computational chemistry analysis in Posit Workbench, BioTeam configured on-demand GPU provisioning tied to active researcher sessions — eliminating idle GPU spend without changing how researchers work. Read the case study
Linux VDI for Scientific Workloads
Many life sciences organizations run scientific applications on Windows environments that were never optimized for computational work. Migrating to Linux-native VDI gives researchers access to the full software stack their work actually requires — without the overhead of maintaining a Windows-based research desktop environment.
How BioTeam does this:
- Assessment of existing Windows/Azure VDI environment and application dependencies
- Linux VDI architecture designed around the specific scientific applications in use
- Migration plan that minimizes researcher disruption during transition
- Configuration of scientific software stack in the new environment
Seen in practice: As part of a broader HPC modernization for a pharmaceutical organization, BioTeam designed and implemented a Linux VDI environment to replace a Windows-based research desktop setup — giving researchers a native environment for their computational chemistry workflows. Read the case study
License-Aware HPC Scheduling
License-aware scheduling isn’t specific to one vendor — it’s a general capability that applies anywhere expensive software licenses need to be coordinated with compute job scheduling. BioTeam has implemented this pattern across multiple tools and environments.
How BioTeam does this:
- License server integrated with Slurm as a native trackable resource
- Scheduler holds jobs until both compute and license tokens are confirmed available
- Works across FlexLM, RLM, and other common license manager types
- Applicable to any computational chemistry, simulation, or licensed scientific software stack
Seen in practice: BioTeam has implemented license-aware scheduling across multiple pharmaceutical engagements involving Schrödinger, Cresset, and other licensed computational tools — in each case, eliminating job failures and idle spend that result from schedulers running blind to license availability.
Read the case studies: Schrödinger at Scale | License-Aware HPC