Venkata Kartik Sadam
Lead Ai Developer @Aviva
Signup · Get unlimited contacts
WORK HISTORY
Lead Ai Developer @Aviva
Norwich, GB
Lead the design and delivery of production GenAI, RAG, and document intelligence systems supporting underwriting and operational workflows.• Architect layout-aware extraction and retrieval pipelines that convert complex documents into structured, grounded responses using Textract, LLMs, and vector-based retrieval.• Own retrieval architecture decisions including document modelling, chunking strategy, grounding reliability, and evaluation approaches to reduce hallucination and improve answer trustworthiness.• Drive the transition from experimental AI solutions to dependable production systems through evaluation-first design, deterministic safeguards, and scalable pipeline architecture.• Establish engineering standards, CI/CD practices, and delivery patterns for AI development, improving maintainability and deployment confidence across the team.• Mentor and guide a team of AI developers, providing technical direction, architectural oversight, and support in navigating complex delivery challenges.
EDUCATION
Sri Venkateswara College of Engineering, Tirupati
Bachelor of Technology - BTech, Electrical and Electronics Engineering
University of East Anglia
MSc Data Science, Computer Science
ABOUT VENKATA KARTIK SADAM
I design and deliver production AI systems that turn complex documents into reliable, decision-ready data.At Aviva, I lead document intelligence and LLM-driven automation across underwriting and operations, transforming manual workflows into scalable, layout-aware extraction pipelines.My focus is production reliability, combining geometry-aware document reconstruction, evaluation-first extraction, and pragmatic LLM architectures that balance capability with deterministic safeguards.Alongside hands-on system design, I guide engineering standards and delivery practices, helping teams move from experimentation to dependable production systems.I’m most interested in building AI that operates under real-world constraints, where accuracy, trust, and maintainability matter more than novelty.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.