Archit Rastogi

Machine Learning Engineer @Pc Cube

Implemented a multilingual (Italian and English) RAG-based document chatbot microservice- Achieved 60% accuracy on a custom completeness metric- Developed a dataset inspired by state-of-the-art research- Leveraged BM25, transformer embeddings, vector databases, Apache Artemis, and LLMs- Saved 20% of users’ time by developing zero-shot and few-shot document classification systems- Utilized RoBERTa and keyword extraction techniques- Supported both Italian and English documents- Reduced manual developer effort by 15–20%- Automated the generation of JUnit tests, integration tests, and JavaDoc documentation- Integrated advanced LLM prompting techniques within the CI/CD pipeline- Developed a time-prediction model for project completion- Used XGBoost and deep learning techniques, achieving 70% accuracy- Enabled effective timeline estimation for project managers and stakeholders- Mentored interns undergoing AI/ML internships- Provided guidance on best practices in model development, data preprocessing, and system integration- Facilitated hands-on experience in real-world AI/ML projects.