# hemantgaikwad06/aimo3-inference-engineering architecture

> Architecture diagram of the GitHub repository hemantgaikwad06/aimo3-inference-engineering, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-09-26). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/hemantgaikwad06/aimo3-inference-engineering
- Repository: https://github.com/hemantgaikwad06/aimo3-inference-engineering

## Overview

This repository records an AIMO3 competition system and its engineering findings; it does not contain the runtime implementation in the supplied tree. The documented workflow runs from a mathematical problem through model inference and verification to an aggregated numerical answer. The graph reflects the README’s described design, not inspected notebook code; deployment and experiment branches are documented separately.

Main entry points: [`docs/inference_pipeline.md`](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/docs/inference_pipeline.md) (Inference workflow), [`README.md`](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md) (Eight attempts), [`docs/lessons_learned.md`](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/docs/lessons_learned.md) (Stochastic variance), [`docs/deployment_failures.md`](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/docs/deployment_failures.md) (LoRA attempts)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_inference["Inference workflow"]
  node_formatting["Harmony formatting<br/>[README.md]"]
  node_orchestrator["Inference workflow"]
  node_vllm["vLLM inference<br/>[README.md]"]
  node_model["GPT-OSS-120B<br/>[README.md]"]
  node_attempts["Eight attempts<br/>[README.md]"]
  node_anchor["Deterministic anchor<br/>[README.md]"]
end

subgraph group_answer["Verification and voting"]
  node_sandbox["Python verification<br/>[README.md]"]
  node_symbolic["Symbolic and numeric checks<br/>[README.md]"]
  node_kernels["Persistent Jupyter kernels<br/>[README.md]"]
  node_consensus["Entropy-weighted voting<br/>[README.md]"]
  node_answer_result["Numerical answer<br/>[README.md]"]
end

subgraph group_engineering["Engineering findings"]
  node_variance["Stochastic variance<br/>[lessons_learned.md]"]
  node_deployment["Deployment constraints"]
  node_finetuning["LoRA attempts"]
  node_experiment_notes["Experiment findings<br/>[notes.md]"]
end

node_user(("Problem source"))

node_user -->|"provides problem"| node_formatting
node_formatting -->|"formats input"| node_orchestrator
node_orchestrator -->|"runs inference"| node_vllm
node_vllm -->|"queries model"| node_model
node_orchestrator -->|"generates attempts"| node_attempts
node_attempts -->|"includes anchor"| node_anchor
node_orchestrator -->|"verifies outputs"| node_sandbox
node_sandbox -->|"uses checks"| node_symbolic
node_sandbox -->|"uses kernels"| node_kernels
node_orchestrator -->|"aggregates attempts"| node_consensus
node_consensus -->|"selects answer"| node_answer_result
node_attempts -->|"exposes variability"| node_variance
node_finetuning -->|"encounters constraints"| node_deployment
node_experiment_notes -->|"records findings"| node_variance

click node_formatting "https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md"
click node_orchestrator "https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/docs/inference_pipeline.md"
click node_vllm "https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md"
click node_model "https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md"
click node_attempts "https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md"
click node_anchor "https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md"
click node_sandbox "https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md"
click node_symbolic "https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md"
click node_kernels "https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md"
click node_consensus "https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md"
click node_answer_result "https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md"
click node_variance "https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/docs/lessons_learned.md"
click node_deployment "https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/docs/deployment_failures.md"
click node_finetuning "https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/docs/deployment_failures.md"
click node_experiment_notes "https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/experiments/notes.md"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_formatting,node_orchestrator,node_vllm,node_model,node_attempts,node_anchor toneBlue
class node_sandbox,node_symbolic,node_kernels,node_consensus,node_answer_result toneAmber
class node_variance,node_deployment,node_finetuning,node_experiment_notes toneMint
class node_user toneIndigo
```

## Components

### Inference workflow

- [Harmony formatting](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md): `README.md`
- [Inference workflow](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/docs/inference_pipeline.md): `docs/inference_pipeline.md`
- [vLLM inference](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md): `README.md`
- [GPT-OSS-120B](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md): `README.md`
- [Eight attempts](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md): `README.md`
- [Deterministic anchor](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md): `README.md`

### Verification and voting

- [Python verification](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md): `README.md`
- [Symbolic and numeric checks](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md): `README.md`
- [Persistent Jupyter kernels](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md): `README.md`
- [Entropy-weighted voting](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md): `README.md`
- [Numerical answer](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/README.md): `README.md`

### Engineering findings

- [Stochastic variance](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/docs/lessons_learned.md): `docs/lessons_learned.md`
- [Deployment constraints](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/docs/deployment_failures.md): `docs/deployment_failures.md`
- [LoRA attempts](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/docs/deployment_failures.md): `docs/deployment_failures.md`
- [Experiment findings](https://github.com/hemantgaikwad06/aimo3-inference-engineering/blob/main/experiments/notes.md): `experiments/notes.md`

### Other components

- **Problem source**

## Connections

- Problem source → Harmony formatting: provides problem
- Harmony formatting → Inference workflow: formats input
- Inference workflow → vLLM inference: runs inference
- vLLM inference → GPT-OSS-120B: queries model
- Inference workflow → Eight attempts: generates attempts
- Eight attempts → Deterministic anchor: includes anchor
- Inference workflow → Python verification: verifies outputs
- Python verification → Symbolic and numeric checks: uses checks
- Python verification → Persistent Jupyter kernels: uses kernels
- Inference workflow → Entropy-weighted voting: aggregates attempts
- Entropy-weighted voting → Numerical answer: selects answer
- Eight attempts → Stochastic variance: exposes variability
- LoRA attempts → Deployment constraints: encounters constraints
- Experiment findings → Stochastic variance: records findings