Akshat Poddar
Computer Science and Mathematics at University of British Columbia
I'm a passionate software engineer who loves building innovative solutions and turning ideas into reality. I specialize in full-stack development with a strong interest in mobile app development and AI/ML Applications. I enjoy working with backend systems to create scalable and efficient applications. I love music, climbing, and movies.
Experience
Software Development Intern
Emozo Labs
- •Developed POC for RAG-based agentic solution to generate actionable insights from structured and unstructured data
- •Used Python to build multi-agent system using LangChain and LangGraph with the latest OpenAI and Ollama LLMs
- •Tried and tested various ML models from HuggingFace for embedding, chunking, and analyzing data
- •Built an end-to-end data ingestion pipeline for pdf, image, docx, and csv files to store in a vector database
- •Built a web application with ReactJS and Tailwind CSS to visualize the data and the insights generated by the agents
Software Development Intern
The Gist: Summarized News App
- • Migrated the existing B2C iOS application from Swift to a modern cross-platform stack using React Native, Expo, andTypeScript, improving development velocity and enabling consistent UI/UX across both iOS and Android platforms
- •Built full-stack features for both B2B and B2C web applications using Next.js and Supabase (PostgreSQL database)
- •Led system design and product features and enforced MVC architecture to maintain a scalable, testable and modular app
- •Collaborated directly with the founder in a fast-paced startup environment, contributing to technical decision-making, product roadmap planning, and long-term scalability strategies
Teaching Assistant (CPSC 317: Computer Networking)
University of British Columbia
- •Led weekly hands-on Java and C tutorials covering HTTP, WebSockets, TCP, DNS, and SMTP protocols
- •Designed lab exercises that guided students through the implementation of networking protocols in Java and C
- •Graded networking assignments and exams, providing targeted feedback to help students improve code quality
Software Engineer Intern
Broadcom Inc.
- •Led the upgrade of Gradle versions, which optimized build processes and reduced overall build times by 20%
- •Performed automated and manual testing on VMs and Docker containers on GCP, managed using Kubernetes
- •Implemented end-to-end CI/CD pipelines using Jenkins, and led the migration of external Java microservices into the main repository, resulting in a 30% reduction in build times
- •Took ownership of improving data flow within the Patch Management System and designed a comprehensive functional specification, which reduced processing delays and improved maintainability
- •Collaborated across AGILE teams to document workflows, deploy microservices, and improve testing coverage
Undergraduate Academic Assistant
University of British Columbia
- •Integrated questions into the PrairieLearn platform to create an open-source Physics problem bank
- •Developed automated testing scripts using PyTest to validate the accuracy and functionality of a new Python package
- •Migrated extensive questions from multiple sources into a unified format using Python and YAML, ensuring data consistency
Projects
Built a cross-platform mobile app using React Native that fetches and displays curated content from the Y-Hacker News API. Designed and implemented a responsive, animated UI with rich media previews using the Microlink API. Integrated ElevenLabs TTS API to convert story text into high-quality speech, enhancing accessibility. Optimized performance by implementing story filtering, pagination, and efficient API data handling.
Developed a machine learning model for semantic segmentation of 3D point clouds to identify and classify object features. Implemented advanced computer vision algorithms using PyTorch and deep learning techniques. The project involved processing large-scale 3D datasets and creating a robust system for real-time object recognition in 3D environments.
Built a comprehensive review platform for JGU (Jindal Global University) that allows students to rate and review courses, professors, and campus facilities. Features include user authentication, review submission, rating systems, and search functionality. The platform helps students make informed decisions about their academic choices.
Developed a real-time messaging app using the MERN stack with TypeScript, featuring Socket.IO for instant communication. Implemented a REST API in Node.js and Express to handle authentication, chatroom routing, and message persistence. Integrated an LLM-powered moderator that enforces on-topic discussions, fact-checks claims, and detects logical fallacies.
Engineered a dynamic cost-calculation app based on user-selected items with advanced filtering and search functionalities. Integrated with Firebase to dynamically fetch product data, utilizing Firestore for data synchronization. Leveraged the MVVM design pattern to ensure efficient data management and separation of concerns. Reduced sales turnaround time by 40% due to improved quote accuracy and faster quote generation.