Ojaswini S.

I am an AI Engineer with 4+ years of experience building and deploying production-ready AI systems across classical machine learning, deep learning, computer vision, NLP, and Generative AI. Unlike many AI developers who focus only on LLMs, I work across the entire AI stack. I believe the best solution isn't always a large language model or an expensive API. Many real-world problems are better solved using classical machine learning or deep learning, resulting in lower infrastructure costs, faster inference, reduced latency, and greater control over your solution. My goal is always to build the most effective system not the most expensive one. Some of the areas I regularly work in include: * Classical Machine Learning (XGBoost, LightGBM, CatBoost, Random Forests, SVMs, feature engineering, predictive modelling, forecasting, anomaly detection, recommendation systems) * Deep Learning (PyTorch, TensorFlow, CNNs, Transformers, Vision Transformers, knowledge distillation, model optimization) * Computer Vision (object detection, image classification, segmentation, OCR, document understanding, face recognition, multi-object tracking, embedding-based search) * NLP & LLMs (RAG, GraphRAG, agentic workflows, fine-tuning, embeddings, semantic search, document QA, information extraction) * Generative AI applications using OpenAI, Anthropic, Gemini, and open-source models * End-to-end AI pipelines from data collection and preprocessing to training, evaluation, deployment, and monitoring I also have extensive experience optimizing AI models for production through knowledge distillation, pruning, quantization, and efficient inference, making models smaller, faster, and more cost-effective for both cloud and edge deployments. On the engineering side, I work comfortably with Python, FastAPI, PostgreSQL, pgvector, asynchronous programming, Docker, GPU acceleration, and cloud deployments. I build complete AI products and APIs that are designed to scale not just research prototypes. Beyond implementation, I enjoy solving difficult research and engineering problems. Whether it's designing a predictive model, improving model accuracy, reducing inference costs, building an intelligent document processing pipeline, or deploying an LLM application, I focus on solutions that are reliable, maintainable, and practical for production. I also lead a team of AI engineers, giving me experience not only in technical execution but also in planning, code quality, mentoring, and delivering projects on time. If you're looking for someone who can understand the problem first, choose the right AI approach, and build a production-ready solution that balances performance, cost, and scalability, I'd be happy to help.