Nizar Karkar

Built and shipped a production AI voice-agent feature for Medicare insurance agencies, integrating LLM tool calling, MCP servers, and RAG knowledge bases. Developed agent tools for callback scheduling, follow-up tasks, and live transfer to human agents.

Automated generation of technical AI labs with a LangGraph workflow, reducing manual creation time by 60%. Designed an agentic AI evaluation workflow and built CI/CD pipelines for AWS Lambda deployments.

Built a RAG-based question-answering system for Order-to-Cash transaction data using LangChain. Implemented hybrid search combining BM25 and semantic vector search, and developed ETL pipelines with AWS Glue and Spark.

Developed data preprocessing pipelines for aircraft sensor data to support predictive maintenance and anomaly detection. Trained and evaluated machine learning models and deployed a Flask REST API for predictions.

Engineering Degree in Computer Science from the National School of Computer Science (ENSI), Tunisia, from September 2021 to June 2024.

Preparatory Cycle in Mathematics and Physics from the Preparatory Institute for Engineering Studies, Tunisia, from September 2019 to June 2021.

Deep Learning Specialization certification from NVIDIA.

Generative AI with Large Language Models certification from DeepLearning.AI.

Data Warehousing Concepts certification from DataCamp.