Headline

AI/ML Engineer | GenAI & RAG Systems | Computer Vision | Building Production-Ready AI Solutions | Ex-Data Science Intern @Digantara

About

Curiosity gets me started, but execution is where I differentiate. I focus on building AI systems that don’t just sit in notebooks they get deployed, used, and deliver measurable impact. My work sits at the intersection of machine learning, deep learning, and Generative AI, with a strong focus on LLMs and RAG-based systems. I’ve built end-to-end pipelines from data preprocessing and feature engineering to model deployment and optimization across domains like healthcare and computer vision, with production-grade reliability in mind. I’m particularly drawn to designing scalable AI systems that are not only performant but also interpretable and robust in real-world environments. Clean architecture, efficient pipelines, and outcome-driven development are non-negotiables in how I build. Currently exploring deeper applications of GenAI in solving complex, unstructured data problems and pushing systems from prototype to production. If you're building in applied AI, GenAI, or data-driven systems, let’s connect!! Always open to meaningful conversations and impactful collaborations.

Location

Gurugram

Work experience

Applied GenAI Engineer at Smile Artists Dental Studio, Oct 2025 - Jan 2026 · 4 mos, Remote, Built an agentic AI system using RAG and LLMs to automate patient query handling and clinical workflows. Designed role-specific AI assistants that processed unstructured medical data (PDF/DOC) into structured insights, improving decision accuracy by 32% and reducing manual effort by 50%. Developed and deployed scalable FastAPI-based microservices with low-latency inference (~1.5s), enabling real-time responses. Focused on embedding-based retrieval, prompt design, and workflow orchestration to create production-ready GenAI systems.

Data Science Intern at Digantara, Nov 2024 - Apr 2025 · 6 mos, Bengaluru, Karnataka, India, Worked on improving how orbital data is modeled and predicted by combining machine learning with physics-based approaches. This significantly improved prediction accuracy (by ~75%) while keeping the system grounded in real-world behavior. Spent a lot of time refining feature engineering and data pipelines, which helped cut down processing time by ~40% and made the system much more efficient. Also built a near real-time pipeline to process incoming data and validate model outputs, so insights could be generated faster and more reliably. A key part of the work involved validating predictions against actual satellite telemetry data—ensuring the models weren’t just accurate in theory, but reliable in practice. Got hands-on experience working with spatial-temporal data and deep learning models, especially in scenarios where both performance and physical consistency mattered. Tech: PyTorch, Python, NumPy, Pandas

Student Intern at Academor, Aug 2023 - Sep 2023 · 2 mos, Remote

Student Intern at Verzeo, Jul 2022 - Aug 2022 · 2 mos, Remote

Education

Bachelor of Technology - B.Tech, Computer Science (Artificial Intelligence and Machine Learning), K.R. Mangalam University

Delhi Public School, Maruti Kunj, Gurgaon

Public links

github: https://github.com/PrachiSwarnim