Jahnvi Sikligar

Contributed to the PDF personalization system, implementing text-personalization modules for dynamically tailored PDFs across multiple user profiles Designed and tested an image replacement algorithm to swap only contextually relevant visuals in PDF outputs using LLM-based content understanding and document editing frameworks Explored applications of Vision-Language Models (VLMs) and multimodal AI pipelines, including extracting spatial relationships from floorplans

Data Scientist at First Living Spaces Pvt. Ltd. (2025-03 – Present)

• Contributed to the PDF personalization system, implementing text-personalization modules for dynamically tailored PDFs across multiple user profiles
• Designed and tested an image replacement algorithm to swap only contextually relevant visuals in PDF outputs using LLM-based content understanding and document editing frameworks
• Explored applications of Vision-Language Models (VLMs) and multimodal AI pipelines, including extracting spatial relationships from floorplans

Group Lead-Data Science at DigiVriddhi (2024-07 – 2024-12)

• Custom-trained machine learning models for object detection, classification, and segmentation, ensuring high model accuracy of above 80% through iterative improvements and data preprocessing.
• Collaborated with cross-functional teams to align data science solutions with business objectives, optimizing existing workflows and contributing to strategic projects.

AI/ML Engineer at Atlas Systems (2023-08 – 2024-05)

• Developed a system using ASR model for audio transcription and custom-trained named entity recognition (NER) with an F1 score of 90%, able to recognise keywords over 20+ fields accurately
• Engineered a RESTful API with Flask to extract structured invoice data from images and PDFs by integrating Azure Cognitive Service's Document Intelligence client library and parsing JSON responses
• Built an AI-powered conversational question-answering system using LangChain to retrieve relevant information from PDFs, leveraging techniques like recursive text splitting, OpenAI embeddings, and FAISS vector stores

Executive Engineer at Raychem RPG Pvt. Ltd. (2019-06 – 2021-08)

• Achieved a 14% increase in bottom-line performance by successfully certifying two product lines in adherence to IEC/EU/UL recognized standards
• Demonstrated regulatory knowledge as the second lead in ATEX audits, resulting in two minor non-conformities

Master's of Science in Artificial Intelligence – Queen Mary University of London (2021-2022)

Bachelor of Instrumentation Engineering in Instrumentation Engineering – University of Mumbai (2015-2019)