Skip to content

docs: add initial project directory and README for task 435#436

Open
Under-the-stars wants to merge 4 commits intomasterfrom
UmdTask435_DATA605_Spring2026_Benchmarking_in_Agentic_Reasoning_for_Data_Science
Open

docs: add initial project directory and README for task 435#436
Under-the-stars wants to merge 4 commits intomasterfrom
UmdTask435_DATA605_Spring2026_Benchmarking_in_Agentic_Reasoning_for_Data_Science

Conversation

@Under-the-stars
Copy link
Copy Markdown
Collaborator

I have added the README file for my project

** Here is the link to my issue ** :
#435 (comment)

@protocorn
Copy link
Copy Markdown
Collaborator

protocorn commented Apr 16, 2026

Please upload the template files in the PR

…chmark

- Implemented AgentState TypedDict with full state management
- Built Planner, EDA Analyst, ML Architect, Quality Reviewer, Report Writer nodes
- Implemented cyclic routing: retry_ml, retry_eda, pass decisions
- Quality Reviewer caught SHAP mismatch and low recall guardrails
- Heart Disease dataset: Accuracy 88.52%, Recall 92.86%, AUC 95.18%
- Added requirements.txt with pinned versions
- Refs #435
@Under-the-stars
Copy link
Copy Markdown
Collaborator Author

Progress Update

Completed:

  • Built complete 5-node LangGraph cyclic agent system
  • Implemented AgentState shared memory across all nodes
  • Quality Reviewer successfully routing with retry_ml and pass decisions
  • Heart Disease benchmark complete

Results:

  • Accuracy: 88.52% | Recall: 92.86% | F1: 88.14% | AUC: 95.18%

Next: NYC Taxi + Amazon Reviews datasets + baseline comparisons

- Added bashrc, copy_docker_files.py, docker_build.version.log
- Added docker_cmd.sh, docker_jupyter.sh, docker_name.sh
- Added etc_sudoers, utils.sh
- Updated Dockerfile and docker_build.sh from template
- Refs #435
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

DATA605_Spring2026_Benchmarking_in_Agentic_Reasoning_for_Data_Science

2 participants