Yogita Adari

Developed ETL pipelines in Python and PostgreSQL, integrating Amadeus and Skyscanner APIs to stream real-time flight data and support synchronous/asynchronous workflows.

Built a flight recommendation engine using GPT-4 and LLM-driven filtering by structuring Amadeus flight data, engineering dynamic prompts with contextual embeddings, and applying multi-dimensional preference modeling, achieving 89% match accuracy in recommendations.

Deployed scalable infrastructure using Google Cloud and Pusher (PaaS), and built Tableau dashboards to track flight traffic and user behavior, enhancing market visibility by 30%.

Collaborated with the founder to shape MVP features, onboard initial paying customers, and reduce time-to-insight by 30% through scalable architecture and CI/CD in a Linux-based Agile environment.

Built a personalized food chatbot using OpenAI API and JSON-based profiles, increasing user demographics by 22% through allergy- and preference-aware recommendations.

Designed ETL pipelines and scraped 20+ restaurant menus using PostgreSQL, NoSQL, and BeautifulSoup, improving data coverage and quality by 35%.

Built a low-cost, high-functioning chatbot using OpenAI’s GPT-4 chatbot, enabling users to instantly place food orders tailored to their preferences, and supporting MVP scalability for Series B funding.

Led the development of a Job application for People with disabilities and members of the LGBTQ+ community for internal and client-facing use at InvisibleStrengths, building responsive UI/UX interfaces using HTML, CSS, JavaScript, and backend workflows that adhered to ADA (HIPAA) and DEIA compliance standards

Integrated Greenhouse and Lever APIs for real-time job matching and developed an admin panel with PostgreSQL and Airtable backend, reducing manual reporting effort by 50%.​

Delivered 4 MVP feature releases in Agile sprints via Jira, enabling the onboarding of 500+ job seekers and 20+ employers during the beta phase.​

Developed a web application for Odyssey Bot, an internal tool designed to streamline remote employee onboarding through integrated Slack and Redmine APIs. Optimized user qualifications using the BERT framework in NLP.

Built ADA-compliant, SEO-optimized, and responsive UI screens using HTML, Semantic UI, CSS, and JavaScript to ensure accessibility and user engagement.

Integrated Slack API for real-time invites and communication setup, and connected to Redmine API to reflect onboarding progress dynamically. Improved onboarding efficiency by 40% and reduced manual admin effort by 60% through automation.