Stratification of Garbage Using Deep Learning
2024
Research on data preprocessing, model development, and performance evaluation for garbage classification using deep learning.
View & download paperOpen to Internships
Hey, I'm Tej 👋
Master's student at San Jose State University
About Me
I'm currently pursuing my Master's in Software Engineering at San José State University, with 2+ years of professional experience building scalable backend systems, APIs, and databases, and collaborating with cross-functional teams to deliver reliable, production-grade software.
During my undergraduate studies, I built a foundation in applied AI and deep learning, including publishing a research paper titled "Stratification of Garbage Using Deep Learning," where I worked on data preprocessing, model development, and performance evaluation.
In my graduate coursework and personal projects, I'm focused on distributed systems and AI agent engineering: orchestrating autonomous task execution across worker nodes, building multi-agent LLM pipelines with LangGraph, and working with retrieval-augmented generation (RAG) to ground LLM responses in real data. My core stack is Python, Java, Go, FastAPI, React, and AWS/Azure. I'm seeking Software Engineer Intern opportunities for Summer 2026 across backend systems, distributed systems, and AI infrastructure teams.
- •Master's in Software Engineering at San José State University with 2+ years of professional experience
- •Research: published paper on "Stratification of Garbage Using Deep Learning" (data preprocessing, model development, performance evaluation)
- •Graduate focus: distributed systems, AI agent orchestration, and RAG pipelines with Python, Java, Go, FastAPI, React, AWS/Azure
- •Seeking Software Engineer Intern opportunities for Summer 2026 (backend systems, distributed systems, AI infrastructure)
Experience & Education
Experience
Software Engineer
Apr 2025 - Jun 2025Deloitte
- •Built and maintained cloud-native backend microservices using Python and FastAPI, supporting production workloads processing 50K+ requests per day
- •Built and enhanced event-driven backend workflows using Kafka and Redis, reducing request latency by 40% and increasing throughput by 3x
- •Developed integrations between modern REST services and legacy IBM mainframe systems, ensuring data consistency across distributed environments
- •Performed production debugging and root-cause analysis for 100+ incidents using structured logging, dashboards, and AI-assisted diagnostic tools, reducing system downtime by 30%
- •Instrumented backend services with metrics and logs and built monitoring dashboards using Prometheus and Grafana, reducing MTTR by 25%
- •Containerized backend services using Docker and automated CI/CD pipelines with Jenkins, reducing operational effort by 60%
- •Worked closely with cross-functional teams including QA, product, and operations to deploy and support reliable backend systems
Software Engineer
Jul 2023 - Mar 2025Carelon Global Solutions
- •Built and developed scalable backend services using Python, FastAPI, Django, and MongoDB supporting enrollment workflows for 1M+ users
- •Developed high-throughput backend data pipelines processing millions of records daily, leveraging AI-assisted validation tools to improve reliability
- •Implemented backend validation and reconciliation logic enforcing 100+ business rules with 99.9% accuracy
- •Built backend aggregation jobs and reporting endpoints feeding internal dashboards monitored via Prometheus and Grafana, improving pipeline issue detection time by 20%
- •Worked with cross-functional engineering and operations teams to operate and support distributed backend systems, contributing to a 30% reduction in production downtime
Software Engineer Intern
Feb 2023 - May 2023Secure Machines
- •Built and developed secure backend-integrated Android applications using Java and Kotlin with AES and SHA-3 encryption
- •Integrated on-device ML inference using TensorFlow Lite, reducing security threats by 60%
- •Developed secure REST API integrations, improving application responsiveness by 40%
Education
Master of Science in Software Engineering
Aug 2025 - May 2027San Jose State University
Relevant Coursework:
Distributed SystemsSystems Software EngineeringData Mining
Bachelor of Engineering in Computer Science
Jul 2019 - May 2023New Horizon College of Engineering
Publications
Projects
Distributed AI Orchestrator
A distributed agent execution system that takes plain-English instructions, uses Gemini to break them into subtasks, and runs them across a fleet of worker nodes. Built the Go API and dispatcher, a Kafka and RabbitMQ pipeline for workload-aware task routing, and a health-checking layer that detects failed workers and automatically reroutes their in-flight work to healthy ones. Instrumented the whole system with Prometheus and Grafana, and packaged it with Docker and Kubernetes manifests for production-style deployment.
GoPythonGemini AIKafkaRabbitMQDockerKubernetes
Access Path AI
A real-time accessible navigation agent. Users describe where they want to go by voice or text, and a LangGraph multi-agent pipeline of three NVIDIA Nemotron models handles speech-to-text, accessibility-aware route reasoning, and a safety check before any route is returned. Built the FastAPI backend and agent orchestration logic, paired with a Next.js and Mapbox frontend for live map rendering. Cut navigation query resolution time by 60% versus a single-model baseline.
LangGraphNVIDIA NemotronFastAPINext.jsMapbox
Cloud Sage
An AI platform that recommends the best cloud provider for a given workload, built with a team as a unified API gateway that compares equivalent services across AWS, GCP, and Azure based on cost and performance. Implemented the FastAPI gateway with JWT auth and rate limiting, cloud provider adapters for all three platforms, and a MongoDB layer for cached pricing and query history, then instrumented it with OpenTelemetry, Prometheus, and Grafana and deployed it on Kubernetes.
FastAPIAWSGCPAzureMongoDBKubernetes
Campus Marketplace
A full-stack marketplace built with a four-person team (Visionary Coders) where students can list, search, and buy/sell items on campus. Built the backend with FastAPI and PostgreSQL and deployed it on Azure, implemented NLP-powered search so listings can be found by natural-language queries instead of exact keyword matches, and added real-time chat between buyers and sellers plus an admin dashboard for moderating listings and users.
Next.jsFastAPIPostgreSQLAzureNLP
SemCacheLM
A quality-aware semantic caching system for LLMs that cuts cost and latency by reusing prior answers when a new query is semantically close enough to trust. Built with a team: an agentic decision layer routes each query to a cache hit, a validation step, or a full LLM fallback, a feedback loop lets users upvote or downvote cached answers to adjust per-entry quality over time, and a lightweight LLM judge catches false hits in the gray zone before reusing an answer. The system is split into FastAPI microservices (gateway, cache, RAG, AI, analytics, orchestrator) backed by Qdrant for vector search and Redis Streams for the analytics pipeline, uses Gemini for embeddings and generation, and ships with Grafana and Prometheus dashboards for observability.
FastAPIQdrantRedisGeminiRAGMicroservices
Garbage Stratification Using Deep Learning
A computer-vision research project that trains a deep learning model to classify waste images into categories (e.g. organic, recyclable, hazardous) to support automated sorting. Covers the full pipeline: data preprocessing and augmentation, model architecture and training, and performance evaluation. Published as a research paper, "Stratification of Garbage Using Deep Learning" (2024).
PythonDeep LearningComputer VisionTensorFlow
Skills
Tap to pause and view
Languages
PythonJavaJavaScriptTypeScriptSQLGoC++PythonJavaJavaScriptTypeScriptSQLGoC++
Web Frameworks
FastAPIDjangoSpring BootREST APIsGraphQLFastAPIDjangoSpring BootREST APIsGraphQL
Cloud & DevOps
AWS (EC2, S3, Lambda)AzureDockerKubernetesCI/CDJenkinsGitPrometheusAWS (EC2, S3, Lambda)AzureDockerKubernetesCI/CDJenkinsGitPrometheus
Data Systems
PostgreSQLMongoDBRedisKafkaPostgreSQLMongoDBRedisKafka
Applied ML
TensorFlowNLPLLMsNumPyPandasMLflowTransformersRAGTensorFlowNLPLLMsNumPyPandasMLflowTransformersRAG