Pavani Rajula
Founder @ NeuCorelytix | Applied AI Engineer | Building Privacy-First & Agentic AI Systems
Hyderabad, India
I'm an AI Engineer, researcher, and founder of NeuCorelytix Solutions, passionate about turning emerging AI capabilities into practical, accessible, and responsible solutions. My work spans Generative AI, agentic systems, LLMs, RAG, privacy-preserving machine learning, and end-to-end AI product development. From building enterprise AI solutions to exploring multimodal applications and autonomous workflows, I enjoy working at the intersection of research, engineering, and real-world problem-solving. My research on privacy-preserving multi-institutional AI for healthcare was accepted as a talk at the PyTorch Conference. More recently, my project Memoiary was recognised as the 1st Runner-up at the Google Cloud Gen AI Academy APAC Cohort 3 Ideathon. Through NeuCorelytix, I'm exploring how we can make AI more sovereign, explainable, affordable, and accessible, especially for communities and businesses that are often overlooked by mainstream technology. I love hackathons, open-source technologies, experimenting with new ideas, and connecting with people who are building meaningful things. Always happy to exchange ideas, collaborate, and learn from fellow builders!
Founder & AI Developer
NeuCorelytix Solutions LLP
- • Founded and leading NeuCorelytix Solutions, focusing on architecting and deploying AI solutions that automate business workflows.
- • Developing and implementing scalable backend systems and AI architectures, enhancing product strategy and driving continuous improvement.
- • Conducting client discovery and requirement engineering, translating complex business needs into production-ready AI solutions with measurable impact.
AI Developer | External Consultant
Data Migration Solutions AG (via NeuCorelytix Solutions LLP)
- • Developed an enterprise RAG-based assistant for conversational retrieval, achieving ~2-3 sec response latency.
- • Created an AI-driven Business Object Proposal workflow utilizing preprocessing and regex-based extraction for heterogeneous data sources.
- • Optimized retrieval pipelines to support complex enterprise queries, maintaining ~5–10 sec end-to-end latency.
Data Scientist / AI Developer
Data Migration Solutions AG
- • Developed an ML-based personal data identification system, achieving ~80% detection performance for sensitive information.
- • Built Text-to-SQL capabilities, allowing business users to generate validated queries in ~10-15 seconds.
- • Created AI-driven data quality enhancement pipelines, resolving ~60% of identified missing values through automated enrichment.
Associate 2 - Developer
StateStreet Corporation
- • Developed and managed regulatory reports using Axiom Controller View, enhancing compliance efficiency.
- • Automated report execution with SQL, UNIX, and shell scripting, streamlining processes.
- • Created Python scripts for data transformation in Hadoop environments, utilizing Hive and PySpark.
- • Contributed to a proof-of-concept project exploring AWS cloud services, including S3 and Lambda.
A multimodal, AI-powered journaling application designed to help people capture, connect, and rediscover meaningful moments through text, voice, images, and video. Built using Gemini on Google Cloud Vertex AI, with intelligent connections between people, places, events, and topics, alongside natural-language memory search. Recognised as the 1st Runner-up at the Google Cloud Gen AI Academy APAC Cohort 3 Ideathon.
An AI-powered software architecture workspace that helps developers explore system designs from multiple SDLC perspectives and visualise architecture using Mermaid diagrams. Supports multiple AI providers, Miro integration, and GitHub Actions to help keep architecture documentation aligned with evolving codebases. Designed to make architectural reasoning more accessible, collaborative, and continuously maintainable.
Privacy-Preserving Multi-Institutional AI for Healthcare
Research exploring privacy-preserving collaborative AI for healthcare, enabling multiple institutions to contribute to machine learning model development without exchanging sensitive patient data. Investigated multi-institutional model fusion and explainability using breast cancer datasets, with a focus on balancing predictive performance, privacy, and interpretability. Research accepted as a talk at the PyTorch Conference.
Masters in Data Science
University of Mannheim
Bachelors of Technology
BVRIT Hyderabad
Attendee · Oct 24, 2026