THARUNIKA VEERARAJ

• Company Overview: Software Company, Chennai
• Developed a Data Extractor Module to process JSON inputs, convert them to JSON-LD, and store structured data in a knowledge base using Semantic Web
• Implemented data mapping and IRI-based entity linking for seamless knowledge ingestion
• Designed a Knowledge Reasoning Module, integrating reasoner calls and SWRL rules using OWLReady2 for dynamic inference
• Executed SPARQL queries for dynamic data aggregation and pushed inferred knowledge to MongoDB
• Built a Data Validation Module to compare inferred knowledge with tabular data, incorporating color-coded mismatches and PDF reporting
• Developed a generic staging module for template mapping and automated data verification
• Software Company, Chennai

Real-time Indian Sign Language for Railway Announcements developed for Indian Railways from Bharathiar University as a project work.

Problem Statement:
AI tool/mobile app for Indian Sign language(ISL) generator from audio-visual content in English/Hindi to ISL
content and vice-versa.
Proposed Methodology:
o Developed a user-friendly web application that translates audio-visual content from English into ISL and vice-versa, enabling smoother communication between hearing individuals and those relying on sign language.
o Created a framework that ensures accuracy and clarity in ISL translation, improving inclusivity in public railway announcements.
o Focused on leveraging computer vision,natural language processing (NLP) and speech recognition to ensure efficient conversion of speech to ISL videos, enhancing user experience and accessibility.

• The current grievance redressal system is limited to text-based submissions in English or Tamil, making it inaccessible to illiterate individuals.
• This project aims to develop a voice-based application powered by LLMs and Generative AI, allowing citizens to effectively report grievances through voice inputs.
• This will enhance accessibility and inclusivity, making the grievance redressal process more user-friendly for everyone.
• Collaborated closely with team members to achieve project objectives and meet deadlines.
• Provided support and guidance to colleagues to maintain a collaborative work environment.

Problem Statement:
Submitted idea to ISRO for developing a context-aware NLP system capable of understanding and retrieving geospatial data based on the context provided in user queries. The proposed system would infer implicit information and deliver relevant geospatial data accordingly.
Proposed Methodology:
o Collected and preprocessed geospatial datasets and a large corpus of geospatial-related texts for training the NLP model.
o Utilized robust frameworks like spaCy and Hugging Face Transformers to train and fine-tune an NLP model tailored to process natural language queries in a geospatial context.
o Integrated the NLP model with a geospatial data retrieval system using PostGIS, developed a user-friendly interface and conducted extensive testing to ensure accuracy, relevance, and a seamless user experience.

Analyzed road accident data from Addis Ababa sub-city police records to uncover trends, hotspots, and contributing factors.
o Deployed an interactive dashboard in Streamlit, allowing users to explore accident patterns and filter data by location, time, and severity.
o Provided actionable insights to assist authorities in improving road safety measures and reducing accident rates in the city.

Responsibilities included preparing the financial statement for FY 2023-24, encompassing subscription fees, employee salaries, routine expenditures, miscellaneous costs, and cash transactions. Utilized Power BI to create a comprehensive annual expenditure report, visually depicting spending patterns. Conducted in-depth analysis and categorization of expenses, identifying significant trends and financial insights. Implemented interactive data exploration tools to empower informed decision-making for financial planning.

• Identified the needs of customers promptly and efficiently.
• Managed time effectively to ensure tasks were completed on schedule and deadlines were met.