Python-Based Anomaly Detection: Finalization/Optimization

Code Finalization and Program Optimization
Job Title:
AI/ML Expert for Finalizing and Optimizing Python-Based Anomaly Detection System

Description:
We are seeking a skilled freelancer with expertise in Python, AI/ML (particularly LSTM and Isolation Forest models), and program optimization to assist in finalizing our anomaly detection system. This project involves critical debugging, model fine-tuning, and simplifying workflows to meet an approaching deadline.

Key Tasks:
1. Code Finalization and Optimization:

Debug and refine critical scripts:
threat_mapping.py: Ensure adversarial testing (e.g., FGSM, PGD, Carlini-Wagner) and anomaly detection are functioning correctly.
evaluation.py: Focus on metrics like precision, recall, F1-score, and latency without retraining models.
hyperparameter_tuning.py: Fine-tune LSTM and Isolation Forest models using efficient hyperparameter optimization techniques.
Ensure seamless integration of pre-trained models with:
New data formats.
Enhanced features from cryptographic operations.
Adaptations to evolving data/testing conditions.
Fine-tune the LSTM model for sequential anomaly detection in SWIFT-like transaction flows and optimize the Isolation Forest model for detecting cryptographic anomalies.
2. Simplifying Program Workflow:

Address non-essential components:
Disable/remove visualizations (e.g., model_visualizer.py) and Elasticsearch integration to streamline functionality.
Simplify program structure:
Ensure scripts such as train_models.py, threat_mapping.py, and evaluation.py can be executed independently in sequence.
Remove reliance on main.py, focusing on a modular execution approach.
Clearly define script dependencies and execution order.
Requirements:
Proficiency in Python and AI/ML model development.
Experience with LSTM networks and Isolation Forests.
Strong debugging skills, particularly in ML pipelines.
Ability to optimize workflows and streamline program structures.
Familiarity with adversarial testing techniques (e.g., FGSM, PGD, Carlini-Wagner) is a plus.
Deliverables:
Fully debugged and optimized scripts (threat_mapping.py, evaluation.py, hyperparameter_tuning.py).
Streamlined program workflow, ensuring independent script execution without reliance on non-essential components or main.py.
Fine-tuned models integrated seamlessly with the existing framework.
Deadline:
This is a time-sensitive project with a strict deadline. Please only apply if you can dedicate immediate attention and complete the tasks promptly.