Define Your Workload.

Configure. Target. Optimize.

Scenarios capture everything about your AI workload: what kind of work it does, how it needs to perform, and what constraints apply. Your scenario becomes context for all recommendations.

lattice.app/scenarios
Lattice Scenario template chooser showing Inference templates with Chat, RAG, Agentic, Code, and Embedding options
Training templates showing Pre-train, LoRA, QLoRA, Full Fine-tuning, RLHF, DPO options

Workload Classification

Define your workload type so Lattice can provide relevant recommendations. From high-volume chat to batch processing, each workload has different optimization priorities.

  • High-Volume Chat: Real-time conversational AI with strict latency
  • RAG Applications: Knowledge base retrieval with accuracy SLOs
  • Code Generation: IDE integration with low-latency completion
  • Agentic Workflows: Multi-step reasoning with tool use
  • Batch Processing: Async workloads with throughput focus
  • Training Jobs: Fine-tuning, RLHF, DPO configurations
Create your first scenario
Scenario configuration form showing SLO settings

SLO Configuration

Set your service level objectives and Lattice will recommend models and configurations that meet your latency, throughput, and availability requirements.

  • P95 Latency: Target response times (e.g., 1000ms)
  • Throughput: Requests per minute requirements
  • Availability: Uptime and reliability targets
  • Token Limits: Input/output token constraints
Configure SLO targets
Scenario template preview showing cost considerations

Budget & Cost Constraints

Define your monthly budget and cost-per-request targets. Lattice factors pricing into recommendations to find the best value for your requirements.

  • Monthly Spend: Budget caps and alerts
  • Cost-per-Request: Per-1K-request targets
  • Cost vs Quality: Tradeoff preferences
  • Reserved vs Spot: Instance pricing options
Set budget constraints
Scenario configuration with compliance and risk profile settings

Compliance & Privacy

Specify compliance requirements like HIPAA, SOC2, and data residency constraints. Lattice filters recommendations to only show compliant options.

  • HIPAA/SOC2: Healthcare and enterprise compliance
  • Data Residency: Geographic constraints for data
  • Audit Requirements: Logging and provenance tracking
  • Privacy Controls: PII handling and anonymization
Compare compliance options

Technical Specifications

Everything you need to know about scenario configuration and workload definition in Lattice.

Workload Categories

  • Inference - Chat, RAG, Code, Embedding
  • Training - SFT, LoRA, RLHF, DPO
  • Comparison - A/B testing scenarios

SLO Parameters

  • Latency - P50, P95, P99 targets
  • Throughput - Requests/min, tokens/sec
  • Availability - Uptime percentages
  • Budget - Monthly and per-request limits

Profile Options

  • Traffic - Low, Medium, High, Burst
  • Risk - Conservative, Balanced, Aggressive
  • Compliance - HIPAA, SOC2, GDPR

Define Your Requirements

Get targeted AI recommendations based on your specific use case and constraints.

Get Lattice for $99