Co-founder @ MagnoliaEd | Research Assistant @ The University of Southern Mississippi
I'm Sugam Panthi, based in Hattiesburg, Mississippi. I have experience building applications using Go and Python. I recently completed an internship as an AI and ML intern @ Prediction3d.
I'm pursuing a Bachelor's in Computer Science at The University of Southern Mississippi.
May 2025 | Hattiesburg, MS Experience
Co-Founder & CTOApr 2025 – Present
MagnoliaEd LLC- Founded an AI-powered EdTech platform delivering personalized instructor chatbots and role-based dashboards, adopted by multiple college classrooms.
- Led full-stack development for a RAG(Retrieval-augmented generation) agentic model, established CI/CD pipelines with GitHub Actions for rapid deployment and automated testing.
- Designed and implemented a Hybrid Monolithic architecture using Docker and Terraform with AWS fargate, enabling independent scaling of chat, analytics, and authentication services, enabling the platform to handle 1,000+ concurrent student sessions.
- Secured $27,000 in startup funding from Co-Builders: powered by Microsoft, attracted angel investors through pitch competitions, and built investor relations.
AI & ML InternMay 2025 – Aug 2025
Prediction 3D- Developed predictive models using PyTorch on multimodal construction site data, by processing images and sensor readings, resulting in improved safety and productivity insights for field teams.
- Integrated AI solutions into web applications with LangChain and Pinecone, collaborating with a 20-engineer team, enabling seamless deployment of ML features to end-users.
- Contributed to the design of a microservice-based data pipeline, separating data ingestion, preprocessing, and model inference for improved maintainability.
- Built Python-based data visualization dashboards by transforming processed datasets into actionable charts, reducing decision-making time for stakeholders by 30%.
Undergraduate Research AssistantNov 2024 - Present
The University of Southern Mississippi- Researched and modeled U.S. plastic waste trends using LSTM networks, cleaning and visualizing large datasets, contributing to two journal paper submissions.
- Designed a React Native recycling logistics app integrating Clerk, MongoDB, and Mapbox, enabling real-time waste transportation tracking for plantations.
- Proposed a modular backend using microservices for data collection, analytics, and reporting, improving scalability for future research projects.
Tech InternFeb 2024 - May 2024
Best One Network- Enhanced the payment confirmation API by transitioning from Django monolith to FlaskWSGI microservice, optimizing processing speed by 20% and improving scalability by 30%.
- Identified, reported, and solved 5 critical bugs during the testing phase, leading to a 10% reduction in post-deployment issues.
- Efficiently integrated and documented the upgraded API into the main codebase using Git and Jira, and contributed to the design of a service-oriented architecture for future integrations.
Skills
Languages
GoPythonJavaScriptTypeScriptSQL
Frameworks & Libraries
ReactNext.jsReact NativeFlaskDjangoLangChainPyTorch
Infrastructure & Databases
DockerTerraformAWS FargateFirebaseMongoDBPineconePostgreSQL
Systems & Protocols
KafkaRaftgRPCRESTMicroservices
Tools & Practices
GitGitHub ActionsCI/CDJiraTurborepo