Varun Tyagi

AI/ML Engineer with hands-on experience in deep learning, NLP, RAG systems, and full-stack development. Proficient in building scalable ML pipelines, training CNNs (ResNet/VGG), fine-tuning LLMs, and deploying vector search solutions. Strong competitive programming background complements data-driven problem-solving and rapid prototyping.

Account Associate – AI Solutions Team at ORANGE BUSINESS (2025-06 – Present)

• Deployed to client's AI engineering team to build RAG pipelines for PDF documents, scraped web data, and speech-to-text transcripts
• Optimized FAISS/pgvector indexes with embedding caching and reranking – reduced query latency by ∼30% while improving top-1 retrieval accuracy
• Designed structured prompt templates and evaluation workflows to ensure reliable, domain-specific LLM responses and reduce hallucinations

AI Engineer Intern at EIGENGRAM (2024-11 – 2025-05)

• Trained CNN-based classifiers (VGG/ResNet/EfficientNet) and tuned hyperparameters to improve accuracy on multiclass image datasets
• Developed feature engineering workflows and compared classical ML models (XGBoost, SVM, Random Forest, KNN, Naïve Bayes) for structured prediction tasks
• Conducted data preprocessing, augmentation, and model evaluation to optimize performance metrics across diverse problem domains

B.Tech. in Electronics & Internet of Things (IoT) – Netaji Subhas University of Technology, New Delhi (2020 – 2024)