Built the full backend development cycle of a persona-driven conversational AI, simulating author-specific GPT-3.5/4 agents with Retrieval-Augmented Generation (RAG)

– Implemented embedding pipelines (OpenAI, Hugging Face) and integrated Pinecone vector DB with persona metadata for low-latency semantic search and scalable top-k retrieval

– Developed scalable chat services for real-time LLM querying, dynamic prompt assembly, automated post-processing, and caching to reduce latency

– Built the MVP with an extensible microservices architecture for independent scaling and features

– Collaborated with data engineering to support iterative testing and tone refinement,

Inventory Management System and Kitchen Tech

• Created scalable backend services and web apps using Kotlin, Kubernetes, Kafka, MySQL, and GraphQL for food production operation management

• Built mobile applications with React Native and Typescript to track inventory items, resulting in reduced inventory loss and food waste.

• Collaborated with multiple products

• Designed and developed backend APIs from scratch to support web and mobile applications using Python and PostgreSQL.

• Implemented fuzzy search functionalities and advanced ranking algorithms with Elastic Search, enhancing search accuracy.

• Led performance optimization, reducing system response time (p99) by 70%, which increased user satisfaction and retention.

• Successfully migrated legacy systems to the cloud with AWS, Docker, and Kubernetes, improving scalability and resource utilization.

• Handled information in a variety of disciplines and researched about the chosen topic.

• Explored different methods of Machine Learning (e.g., reinforcement learning, Bayesian methods, Support Vector Machines, and so on).

• Focused on classification algorithms.