Check K.

- Deployment & Integration:AI models (NLP, multimodal LLMs, SLMs, computer vision) with ONNX Runtime, LangChain, MLflow, MLLM-Tool.

- Maintenance & Testing:LLM deployment tools, load, stress, and endurance testing.

- Monitoring & Quality:Performance, scalability, and reliability tracking.

- Optimization:Improvement via RAG, fine-tuning, and the LLM as a Judge approach.

- Innovation & Research:Experimentation (LLM, SLM, CNN, computer vision).

- Production deployment of ML modelsfor airport staff optimization (security control posts, resource allocation).

- Automated pipelines on GCPwith Vertex AI Pipelines, Cloud Build, Cloud Run, Terraform, and GitLab CI.

- Model Monitoring and Retrainingwith MLflow, Evidently AI, and Vertex AI Monitoring; integration of drift/OOD detection (Jensen-Shannon divergence, KL, Kolmogorov-Smirnov test, MAPIE) across various time horizons and operational assets.

- Containerization & Orchestrationwith Docker, Kubernetes, Airflow, and Argo for scalable workflows.

- Real-time dashboardsvia BigQuery, Streamlit, and Cloud Functions to track performance, data quality, and latency.

Production deployment of trustworthy AI with several partners

-End-to-end development in Machine Learning & Deep Learning(designing models from scratch)

- Deployment and production of models(Edge AI compatibility) with a strong focus on robustness, explainability, and reliability

- Design and implementation of pipelinesfor data processing, training, evaluation, and retraining

- Evaluation and improvement of the technology stack:code quality, PEP8 compliance, technical documentation, CI/CD best practices

- Monitoring of model drift in production(data drift, concept drift, performance)

- Design and implementation of Big Data architecturesfor scalable and resilient systems

- Preprocessing and preparationof data for training and fine-tuning of large language models (LLMs)

- Design of RAG systems(Retrieval-Augmented Generation): data indexing, semantic search, vectorization, query orchestration, and response quality improvement

- Use of advanced LLM tools such as LangChain(and associated frameworks) for orchestrating prompt chains, agents, and external tools

- Expertise in contextual embeddingsof words and sentences with LLMs like BERT, GPT-3, and GPT-4

- In-depth experience with Transformers and LLMsfor various NLP use cases: text generation, semantic search, question-answering, sentiment analysis