I'm a full-stack and AI/ML engineer with 10+ years of experience building production systems across healthcare, insurance, and finance. I specialize in architecting end-to-end solutions — from designing scalable backend APIs and modern frontend experiences to integrating large language models into real business workflows. I've led projects that reduced codebases by 35%, cut dependency overhead by 43%, and brought AI-powered tools directly into the hands of entire organizations. I hold a Master's in Computer Science from the University of Texas at Arlington.
Architected and evolved ZanderGPT, an internal AI assistant that transformed how the company accesses insurance knowledge. Migrated from a custom open-source LLM setup to AWS Bedrock Knowledge Bases for more reliable, scalable retrieval. Implemented streaming responses for a significantly faster user experience. Integrated directly into Microsoft Teams — removing the friction of navigating to a separate URL and putting AI assistance where the entire company already works.
Built an AI-powered customer simulation tool to help agents practice real-world insurance conversations. Developed a FastAPI backend connected to AWS Bedrock for intelligent, context-aware responses. Added full voice interaction — users can speak directly in the chat interface with speech-to-text capture, and responses are converted back to speech natively in the browser, creating a hands-free conversational experience.
Replaced a cumbersome workflow — previously, finding a recorded call meant navigating date-specific folders, searching by phone number, downloading an MP3, and playing it locally. Call Explorer lets users search any call instantly and play it directly in the browser. Pulls recordings from S3, fed by the Call Recordings Archive CRON pipeline.
Led a comprehensive overhaul of the Zander marketing website, upgrading Next.js and migrating the entire codebase from JavaScript to TypeScript. Delivered measurable improvements in long-term maintainability.
Extended the Main-API with new third-party integrations including Pure Caller ID. Introduced a bearer token architecture enabling external systems to push notifications securely and directly to the database. Led an overhaul of the Call Recordings Archive system.
Developed two production API endpoints: one to retrieve documents from Epic Systems, converting byte-format responses into usable documents; another to predict health class based on patient attributes (height, weight, age, sex) to support underwriting workflows.
Migrated both Reliaquote and Reliashield from a legacy CraftCMS setup to a modern Next.js framework. Set up dedicated headless WordPress instances for blog content on both platforms and integrated them seamlessly with the Next.js projects.
Designed and implemented a cloud-based Electronic Health Record system for an in-house surgery center using Agile methodology. Integrated with the Allscripts Patient Management system. Built an autocomplete search feature for ICD-10 codes powered by Elasticsearch.
Applied machine learning and deep learning to build applications aimed at making medical workflows more efficient. Implemented text generation using Google's BERT model for use within the EHR system. Performed network penetration testing using Kali Linux.
Built a cloud-based dynamic report generation application for the billing department. Developed a medication inventory system. Created a clinical trial Electronic Data Capture system using Waterfall methodology.
Maintained and enhanced retail management software built on a modular, plug-and-play ABAP architecture, allowing flexible customization for enterprise clients. Worked within a SCRUM team on a Java and J2EE-based platform.
Designed and implemented a system to monitor all chemicals and sterilization equipment at the Nanotechnology Research Center.
Maintained and enhanced Summit FT, a finance software system. Led the migration of the Retail Enterprise Suite from Oracle Forms to a modern cloud-based architecture. Worked within a team alongside developers, QA, and project managers.
Open Source · Under Development
An open-source inventory management system for small retail shops, online sellers, and manufacturers. Mobile-first, simple, and free to self-host. Features include product management, stock tracking, barcode scanning, and real-time updates.
npm Package · Monthly Updates
Fast, lightweight US ZIP code lookup library with zero runtime dependencies. Supports lookups by ZIP code, city, county, and geographic radius. Works universally across Node.js, React, Next.js, and browsers.
Karthik Gangadharaiah · Full-Stack · AI/ML · 10+ Years