Design space exploration

For ONNX & Runtime Engines, tradeoffs are accuracy, latency, power, and workflow risk.

Option A — conservative

• Conservative baseline: helps predictable operation

• Risk: lower peak performance

• Validate with: first deployment

Option B — balanced

• Balanced optimization: helps good efficiency

• Risk: needs tuning discipline

• Validate with: mainline production

Option C — aggressive optimization

• Aggressive optimization: helps best resource use

• Risk: higher fragility

• Validate with: mature pipelines

Option D — architecture refactor

• Architecture refactor: helps long-term scalability

• Risk: schedule risk

• Validate with: repeated bottlenecks

DESIGN SPACE — ONNX & Runtime Engines
accuracy <-> latency <-> power <-> engineering risk

Design pitfalls

• Optimizing one KPI without system-level impact

• Skipping fallback plan for low-confidence regions

Tradeoff curve

BEFORE / AFTER — ONNX & Runtime Engines

metric quality
  ^
  |                        o target region
  |                 o post-fix + regression
  |            o
  |      o baseline failing run
  +-----------------------------------------> iteration
      evidence audit   fix    full validation

AI-VLSI deep dive

Deployment quality is determined by compatibility, observability, and rollback discipline.

Concept diagram

DEPLOYMENT LOOP

model export -> runtime integration -> production monitor -> retrain/rollback

Metric graph

DEPLOYMENT HEALTH

conversion pass rate    ████████
latency SLA pass        ███████
drift incidents         ██

Reports and artifacts

• ONNX conversion report

• quantized eval sheet

• stack latency breakdown

• drift incident tracker

Mini case study

A runtime operator gap was caught pre-release by compatibility checks and shadow deployment.

Debug branches

• Validate operator coverage

• Trace firmware-driver latency

• Check drift thresholds and rollback path

Senior review question

Ask: what evidence connects this ML claim to a concrete VLSI workflow decision and owner signoff?

Key takeaways

• Every AI claim should map to a measurable engineering outcome.

• Validate both model quality and hardware/runtime feasibility before adoption.

Common pitfalls

• Optimizing benchmark metrics that do not correlate with signoff goals.

• Ignoring data drift and calibration after deployment.

• Shipping ML workflows without clear rollback ownership.

Execution drill pack 1

Use this pack to rehearse AI-for-VLSI decision making on ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space: metric framing, mechanism proof, hardware implications, and release safety.

Evidence checklist

• Metric context includes workload, dataset slice, and revision tags.

• Mechanism explanation links model behavior to observed outcome.

• Hardware/runtime feasibility is profiled, not assumed.

• Owner and rollback path are documented before rollout.

Review prompts

1. Which decision will this model output influence?

2. What is the first failing layer when metric regresses?

3. Which owner applies the smallest reversible fix?

4. What validation matrix is required before deployment?

Evidence capsule

AI-VLSI EVIDENCE CAPSULE 1

PATH: ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space
WORKLOAD SLICE: <name>
PRIMARY METRIC: <value/trend>
FIRST FAILING LAYER: <data/model/runtime/hardware>
OWNER: <name>
PRIMARY ARTIFACT: <report/profile/dashboard>
DECISION: <ship / rollback / escalate>

Execution drill pack 2

Use this pack to rehearse AI-for-VLSI decision making on ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space: metric framing, mechanism proof, hardware implications, and release safety.

Evidence checklist

• Metric context includes workload, dataset slice, and revision tags.

• Mechanism explanation links model behavior to observed outcome.

• Hardware/runtime feasibility is profiled, not assumed.

• Owner and rollback path are documented before rollout.

Review prompts

1. Which decision will this model output influence?

2. What is the first failing layer when metric regresses?

3. Which owner applies the smallest reversible fix?

4. What validation matrix is required before deployment?

Evidence capsule

AI-VLSI EVIDENCE CAPSULE 2

PATH: ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space
WORKLOAD SLICE: <name>
PRIMARY METRIC: <value/trend>
FIRST FAILING LAYER: <data/model/runtime/hardware>
OWNER: <name>
PRIMARY ARTIFACT: <report/profile/dashboard>
DECISION: <ship / rollback / escalate>

Execution drill pack 3

Use this pack to rehearse AI-for-VLSI decision making on ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space: metric framing, mechanism proof, hardware implications, and release safety.

Evidence checklist

• Metric context includes workload, dataset slice, and revision tags.

• Mechanism explanation links model behavior to observed outcome.

• Hardware/runtime feasibility is profiled, not assumed.

• Owner and rollback path are documented before rollout.

Review prompts

1. Which decision will this model output influence?

2. What is the first failing layer when metric regresses?

3. Which owner applies the smallest reversible fix?

4. What validation matrix is required before deployment?

Evidence capsule

AI-VLSI EVIDENCE CAPSULE 3

PATH: ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space
WORKLOAD SLICE: <name>
PRIMARY METRIC: <value/trend>
FIRST FAILING LAYER: <data/model/runtime/hardware>
OWNER: <name>
PRIMARY ARTIFACT: <report/profile/dashboard>
DECISION: <ship / rollback / escalate>

Execution drill pack 4

Use this pack to rehearse AI-for-VLSI decision making on ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space: metric framing, mechanism proof, hardware implications, and release safety.

Evidence checklist

• Metric context includes workload, dataset slice, and revision tags.

• Mechanism explanation links model behavior to observed outcome.

• Hardware/runtime feasibility is profiled, not assumed.

• Owner and rollback path are documented before rollout.

Review prompts

1. Which decision will this model output influence?

2. What is the first failing layer when metric regresses?

3. Which owner applies the smallest reversible fix?

4. What validation matrix is required before deployment?

Evidence capsule

AI-VLSI EVIDENCE CAPSULE 4

PATH: ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space
WORKLOAD SLICE: <name>
PRIMARY METRIC: <value/trend>
FIRST FAILING LAYER: <data/model/runtime/hardware>
OWNER: <name>
PRIMARY ARTIFACT: <report/profile/dashboard>
DECISION: <ship / rollback / escalate>

Execution drill pack 5

Use this pack to rehearse AI-for-VLSI decision making on ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space: metric framing, mechanism proof, hardware implications, and release safety.

Evidence checklist

• Metric context includes workload, dataset slice, and revision tags.

• Mechanism explanation links model behavior to observed outcome.

• Hardware/runtime feasibility is profiled, not assumed.

• Owner and rollback path are documented before rollout.

Review prompts

1. Which decision will this model output influence?

2. What is the first failing layer when metric regresses?

3. Which owner applies the smallest reversible fix?

4. What validation matrix is required before deployment?

Evidence capsule

AI-VLSI EVIDENCE CAPSULE 5

PATH: ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space
WORKLOAD SLICE: <name>
PRIMARY METRIC: <value/trend>
FIRST FAILING LAYER: <data/model/runtime/hardware>
OWNER: <name>
PRIMARY ARTIFACT: <report/profile/dashboard>
DECISION: <ship / rollback / escalate>

Execution drill pack 6

Use this pack to rehearse AI-for-VLSI decision making on ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space: metric framing, mechanism proof, hardware implications, and release safety.

Evidence checklist

• Metric context includes workload, dataset slice, and revision tags.

• Mechanism explanation links model behavior to observed outcome.

• Hardware/runtime feasibility is profiled, not assumed.

• Owner and rollback path are documented before rollout.

Review prompts

1. Which decision will this model output influence?

2. What is the first failing layer when metric regresses?

3. Which owner applies the smallest reversible fix?

4. What validation matrix is required before deployment?

Evidence capsule

AI-VLSI EVIDENCE CAPSULE 6

PATH: ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space
WORKLOAD SLICE: <name>
PRIMARY METRIC: <value/trend>
FIRST FAILING LAYER: <data/model/runtime/hardware>
OWNER: <name>
PRIMARY ARTIFACT: <report/profile/dashboard>
DECISION: <ship / rollback / escalate>

Execution drill pack 7

Use this pack to rehearse AI-for-VLSI decision making on ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space: metric framing, mechanism proof, hardware implications, and release safety.

Evidence checklist

• Metric context includes workload, dataset slice, and revision tags.

• Mechanism explanation links model behavior to observed outcome.

• Hardware/runtime feasibility is profiled, not assumed.

• Owner and rollback path are documented before rollout.

Review prompts

1. Which decision will this model output influence?

2. What is the first failing layer when metric regresses?

3. Which owner applies the smallest reversible fix?

4. What validation matrix is required before deployment?

Evidence capsule

AI-VLSI EVIDENCE CAPSULE 7

PATH: ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space
WORKLOAD SLICE: <name>
PRIMARY METRIC: <value/trend>
FIRST FAILING LAYER: <data/model/runtime/hardware>
OWNER: <name>
PRIMARY ARTIFACT: <report/profile/dashboard>
DECISION: <ship / rollback / escalate>

Execution drill pack 8

Use this pack to rehearse AI-for-VLSI decision making on ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space: metric framing, mechanism proof, hardware implications, and release safety.

Evidence checklist

• Metric context includes workload, dataset slice, and revision tags.

• Mechanism explanation links model behavior to observed outcome.

• Hardware/runtime feasibility is profiled, not assumed.

• Owner and rollback path are documented before rollout.

Review prompts

1. Which decision will this model output influence?

2. What is the first failing layer when metric regresses?

3. Which owner applies the smallest reversible fix?

4. What validation matrix is required before deployment?

Evidence capsule

AI-VLSI EVIDENCE CAPSULE 8

PATH: ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space
WORKLOAD SLICE: <name>
PRIMARY METRIC: <value/trend>
FIRST FAILING LAYER: <data/model/runtime/hardware>
OWNER: <name>
PRIMARY ARTIFACT: <report/profile/dashboard>
DECISION: <ship / rollback / escalate>

Execution drill pack 9

Use this pack to rehearse AI-for-VLSI decision making on ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space: metric framing, mechanism proof, hardware implications, and release safety.

Evidence checklist

• Metric context includes workload, dataset slice, and revision tags.

• Mechanism explanation links model behavior to observed outcome.

• Hardware/runtime feasibility is profiled, not assumed.

• Owner and rollback path are documented before rollout.

Review prompts

1. Which decision will this model output influence?

2. What is the first failing layer when metric regresses?

3. Which owner applies the smallest reversible fix?

4. What validation matrix is required before deployment?

Evidence capsule

AI-VLSI EVIDENCE CAPSULE 9

PATH: ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space
WORKLOAD SLICE: <name>
PRIMARY METRIC: <value/trend>
FIRST FAILING LAYER: <data/model/runtime/hardware>
OWNER: <name>
PRIMARY ARTIFACT: <report/profile/dashboard>
DECISION: <ship / rollback / escalate>

Execution drill pack 10

Use this pack to rehearse AI-for-VLSI decision making on ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space: metric framing, mechanism proof, hardware implications, and release safety.

Evidence checklist

• Metric context includes workload, dataset slice, and revision tags.

• Mechanism explanation links model behavior to observed outcome.

• Hardware/runtime feasibility is profiled, not assumed.

• Owner and rollback path are documented before rollout.

Review prompts

1. Which decision will this model output influence?

2. What is the first failing layer when metric regresses?

3. Which owner applies the smallest reversible fix?

4. What validation matrix is required before deployment?

Evidence capsule

AI-VLSI EVIDENCE CAPSULE 10

PATH: ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space
WORKLOAD SLICE: <name>
PRIMARY METRIC: <value/trend>
FIRST FAILING LAYER: <data/model/runtime/hardware>
OWNER: <name>
PRIMARY ARTIFACT: <report/profile/dashboard>
DECISION: <ship / rollback / escalate>

Execution drill pack 11

Use this pack to rehearse AI-for-VLSI decision making on ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space: metric framing, mechanism proof, hardware implications, and release safety.

Evidence checklist

• Metric context includes workload, dataset slice, and revision tags.

• Mechanism explanation links model behavior to observed outcome.

• Hardware/runtime feasibility is profiled, not assumed.

• Owner and rollback path are documented before rollout.

Review prompts

1. Which decision will this model output influence?

2. What is the first failing layer when metric regresses?

3. Which owner applies the smallest reversible fix?

4. What validation matrix is required before deployment?

Evidence capsule

AI-VLSI EVIDENCE CAPSULE 11

PATH: ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space
WORKLOAD SLICE: <name>
PRIMARY METRIC: <value/trend>
FIRST FAILING LAYER: <data/model/runtime/hardware>
OWNER: <name>
PRIMARY ARTIFACT: <report/profile/dashboard>
DECISION: <ship / rollback / escalate>

Execution drill pack 12

Use this pack to rehearse AI-for-VLSI decision making on ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space: metric framing, mechanism proof, hardware implications, and release safety.

Evidence checklist

• Metric context includes workload, dataset slice, and revision tags.

• Mechanism explanation links model behavior to observed outcome.

• Hardware/runtime feasibility is profiled, not assumed.

• Owner and rollback path are documented before rollout.

Review prompts

1. Which decision will this model output influence?

2. What is the first failing layer when metric regresses?

3. Which owner applies the smallest reversible fix?

4. What validation matrix is required before deployment?

Evidence capsule

AI-VLSI EVIDENCE CAPSULE 12

PATH: ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space
WORKLOAD SLICE: <name>
PRIMARY METRIC: <value/trend>
FIRST FAILING LAYER: <data/model/runtime/hardware>
OWNER: <name>
PRIMARY ARTIFACT: <report/profile/dashboard>
DECISION: <ship / rollback / escalate>

Execution drill pack 13

Use this pack to rehearse AI-for-VLSI decision making on ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space: metric framing, mechanism proof, hardware implications, and release safety.

Evidence checklist

• Metric context includes workload, dataset slice, and revision tags.

• Mechanism explanation links model behavior to observed outcome.

• Hardware/runtime feasibility is profiled, not assumed.

• Owner and rollback path are documented before rollout.

Review prompts

1. Which decision will this model output influence?

2. What is the first failing layer when metric regresses?

3. Which owner applies the smallest reversible fix?

4. What validation matrix is required before deployment?

Evidence capsule

AI-VLSI EVIDENCE CAPSULE 13

PATH: ai-vlsi/edge-deployment-mlops/onnx-and-runtime-engines/design-space
WORKLOAD SLICE: <name>
PRIMARY METRIC: <value/trend>
FIRST FAILING LAYER: <data/model/runtime/hardware>
OWNER: <name>
PRIMARY ARTIFACT: <report/profile/dashboard>
DECISION: <ship / rollback / escalate>

Principal AI-VLSI review addendum

Standard model interchange and runtime engines bridge training frameworks to production execution stacks on edge and embedded targets.

Metric: model conversion success rate, runtime compatibility, and inference latency