Suvojit Hore

Senior AI Engineer working for a leading Australian insurance client in building and productionizing AI solutions to automate insurance workflows.

• Developed and productionized an agentic claims-processing system that autonomously analyzes incidents, maintains memory, and drives end-to-end claims decisions with Claude Opus.

• Built a PII deidentification tool leveraging LLMs with iterative feedback loops to mask and replace sensitive data.

• Deployed a GenAI chatbot enabling stakeholders to query insurance documents, retrieve policy details, and file claims - OpenAI GPT-5, Mistral.

• Built an application to create AI Twins, digital avatars that replicate an individual's appearance, voice, and speech - Agentic AI, RAG, lip-sync.

Gen AI Developer working for a leading US client automating image, text, and audio compliance using multimodal large language models.

• Compliance solution for advertisements - Using multimodal LLMs to verify image, video, audio, and text data in ads to ensure they follow guidelines and rules.

• Chatbot with open-source LLM finetuned to answer user queries about the new legal framework of India - Bhartiya Nyaya Sanhita.

Senior Data Scientist working with US Laptop and Hardware companies to automate advertisement compliance using Generative AI.

• Built multimodal LLM-based compliance solution to verify image, video, audio, and text data in advertisements to ensure they follow guidelines and rules for computer hardware clients.

• Developed a chatbot with a fine-tuned open-source LLM to answer queries about India’s new legal framework - Bhartiya Nyaya Sanhita.

Senior Data Scientist working with some of the largest retailers in UK, EU and around the world to solve business problems with data science, generative AI and analytics.

Domains - Generative AI, NLP, Computer Vision.

Projects -

• Chatbot for business users using open source, finetuned Large Language Models to convert Natural Language to SQL Queries and Python code and provide data insights.

• AI dashboard and bot to assist data scientists and engineers with coding, error debugging, and answering internal queries.

• Recommendation Engine for new products for the UK’s Largest retailer.

• Healthy food recommendation using XgBoost and Fasttext for leading Norwegian clients.

• Media campaign evaluation and control matching using Random Forest and K-means Clustering.

• Customer family segmentation using XgBoost Classifier

Research Data Scientist working as part of a product team in the R&D unit of Infosys. Primary Domains - NLP, Computer Vision.

Key Projects:

• Smart Chatbot - Built and productionized a rule-based chatbot for a British telecom company using Dialogflow, which is integrated with Confluence, SharePoint, Google Data Catalog, and BigQuery.

• Cluster Analysis - Developed an automated cluster analysis dashboard for large unstructured datasets with just a few clicks.

• Real-Time Image Classification - Trained and deployed a CNN model for meat quality classification in a production line for a major US food processing company.

• Synthetic Image Generation - Created a pix2pix GAN-based image generation service for creating synthetic images, customizable for client needs.

• Data Privacy Chatbot - Built an interactive chatbot using Rasa as a helper for the product - IEDPS.

• Form OCR and Generation - Trained and deployed a form OCR service using neural networks that can read new forms, and also generate identical forms filled with printed as well as handwritten text.

Thumbnail for Insta Award - Infosys