Led the design and deployment of a retrieval-augmented generation (RAG) driven knowledge base for internal support. Consolidated and ingested external and internal documentation from GitHub, SharePoint, DOORS, Confluence, and Jira into a centralized index using a custom data pipeline and LlamaIndex. Built a chat bot for internal ticketing that autonomously answers questions by drawing on structured sources, ensuring accuracy via pre-generation checks. Defined and implemented domain-specific retrieval logic, embeddings, and cross-functional integrations with AI-powered assistants. Drove improvements in run time, response accuracy, and consistency through iterative testing, log analysis, and collaboration with software, hardware, and data engineering teams. Optimized operational support by reducing ticket volume and manual workload while enabling scalable knowledge access across teams. Planned and evaluated deployment options on AWS Bedrock, including hosting, orchestration, and cost/per

Software Engineer Intern at Continental Serbia, Novi Sad

April 1, 2023 - November 1, 2023

Developed software components in the Application Layer of AUTOSAR, following automotive standards and system requirements. Performed log analysis, debugging, and issue resolution in collaboration with cross-functional software and hardware teams.

Software Engineer Intern at Continental Serbia, Novi Sad

June 1, 2022 - August 1, 2022

Developed a domain-specific insurance chatbot using RNN architectures to answer user queries. Implemented two approaches: a high-precision model using Pandas, NumPy, Keras, NLTK, Seaborn, and Matplotlib for data preprocessing and model training; and a high-recall model leveraging pretrained word vectors and multiple similarity metrics (e.g., cosine similarity) to maximize coverage. Performed dataset cleaning, tokenization, lemmatization, exploratory analysis, and testing to optimize model accuracy. Tuned hyperparameters, evaluated model performance, and improved answer quality through iterative experimentation. Built a recommendation system to suggest the most suitable insurance policies for each user. Designed feature-engineering pipelines and selected appropriate machine-learning algorithms.