Headline

GenAI/ML @ Ascendion | IIT Madras

About

Data Professional with expertise in gathering, cleaning, and organizing data for both technical and non-technical stakeholders. Proficient in Predictive Modeling, Machine Learning, Deep Learning and Generative AI, demonstrating a comprehensive understanding of Data Science. Proficient in deploying ML pipelines and models to production. Highly organised, motivated, and diligent, leveraging a strong background in analytics to drive insightful and actionable results. Committed to applying my skills and knowledge to contribute to the success of a company's mission.

Location

Pune District

AI stack

codex

opencode

Work experience

GenAI/ML Engineer at Ascendion, Jul 2024 - Present · 2 yrs 1 mo, Pune, Maharashtra, India, Developed a high-precision PDF table extraction pipeline using Microsoft’s Table Transformer and CRAFT OCR, achieving ~95% structural accuracy on complex, pivot-style financial tables. Designed and deployed a scalable Retrieval-Augmented Generation (RAG) platform to surface critical policy insights from dense financial disclosure bundles (e.g., Mastercard, VISA). The solution integrated a Neo4j Knowledge Graph via LightRAG for semantic reasoning, a high-throughput Haystack retriever with cross-encoder reranking, and a prompt-optimized GPT-4 model — resulting in 4× faster insight generation with highly structured, context-aware outputs. Co-developed resilient FastAPI microservices as part of a cross-functional offshore team for a multi-intent IVR system handling balance and statement inquiries. The system streamed Gemini Flash 2.0 responses via Vertex AI and Neo4j for real-time user financial attribute context, achieving >90% classification accuracy and sub-5s latency in Production.

Education

Bachelor of Science - BS, Data Science and Applications, Indian Institute of Technology, Madras

Bachelor of Science - BS, Electronics and Communication Technology, Chatrapati Sahuji Maharaj Kanpur University, Kanpur