All projects

Agentic AI & MLOps / 3D.S.F

AgenticIA · Multi-agent ML / LLM automation

A multi-agent platform connecting ingestion, training, evaluation, deployment and monitoring.

My role
Internship · Agentic AI & MLOps
Period
Jan — May 2026

Inside the system

Connect the model lifecycle

A conceptual overview of the AgenticIA workflow described in my CV.

  1. 01

    Ingest

    Prepare and version data

  2. 02

    Train & tune

    Experiments and HPO

  3. 03

    Evaluate & deploy

    Model evaluation and serving

  4. 04

    Monitor

    Performance and drift

  • Orchestration

    LangGraph coordinates the workflow

  • Feedback

    Diagnose → corrective action → retrain

01 / The context

What the project addresses

ML and LLM workflows involve many connected steps and tools. AgenticIA brings them together through a multi-agent automation platform.

02 / My contribution

What I worked on

Designed and developed the platform, integrated lifecycle tools and implemented an adaptive monitoring and corrective-action loop.

The challenge

Coordinate the model lifecycle and connect performance monitoring to actions such as diagnosis and retraining.

03 / The approach

How the pieces fit together

  1. Orchestrate ingestion, training, hyperparameter optimization and evaluation with LangGraph.
  2. Integrate MLflow and DVC for experiment tracking and versioning, and Optuna / Ray Tune for optimization.
  3. Connect FastAPI, Docker and Kubernetes with deployment workflows, plus vLLM and LangSmith for LLM workflows.
  4. Use monitoring and drift detection to let agents diagnose degradation and trigger corrective actions.

04 / The outcome

What the work delivered

  • Connected ML / LLM lifecycle stages in one agentic workflow.
  • Implemented a feedback loop for monitoring, diagnosis and corrective action.
Next case study

A second life for images