Arkajit Dutta
San Diego,CA
Summary
Computer Science Master’s student specializing in AI and real-time systems, with 3+ years of industry experience delivering scalable software solutions. Built deep learning and computer vision systems in academia while engineering high-performance enterprise software at GUESS Inc. Proficient in Java, Python, C++, CUDA, and AWS. Versatile across research and production, with a proven track record of shipping reliable, high-traffic applications.
Overview
4
4
years of professional experience
Work History
Applied AI Research Assistant
San Diego State University
08.2024 - Current
- Developed a computer vision-based navigation system for drones, integrating deep learning techniques for real-time obstacle detection and path planning in confined indoor spaces
- Implemented NLP-based voice command recognition, leveraging LangChain and LlamaIndex to enable intuitive drone control and improve real-time responsiveness
- Trained and fine-tuned GPT-4 on AWS P4D instances using PyTorch, optimizing model parameters and enhancing retrieval-augmented decision-making, resulting in a 40% improvement in path success rate across 100+ test runs
- Enhanced drone path planning efficiency by constructing a knowledge graph that captured environmental constraints and drone capabilities reducing navigation errors by 30% and optimizing flight efficiency
Project Lead - Crowd Flow Simulation
SDSU Research
12.2024 - 05.2025
- Led the design of a hybrid deep learning pipeline for crowd trajectory forecasting, combining ConvLSTM and Social-LSTM models trained on 10,000+ frame sequences with 500k+ annotated trajectories
- Incorporated Trajectory Transformer to model global attention across pedestrian paths, outperforming RNN-based baselines on long-horizon predictions
- Accelerated real-time inference through TorchScript compilation and CUDA kernel tuning, sustaining sub-100ms latency for simulation integration
- Engineered real-time visualization overlays using OpenCV, enabling sub-100ms inference latency and predictive simulation of future crowd flows up to 10 seconds ahead
Java Full Stack Developer
GUESS Inc.
08.2021 - 07.2024
- Perform full life-cycle development, including requirements analysis, requirements decomposition, system design, and unit integration on Enovia PLM system using Java, JavaScript, CSS
- Created 5+ materialized views for data modeling and replication in Oracle SQL, reducing query time by 15% and optimizing system response rates
- Integrated the application with ERP systems to facilitate internal data sharing and purchase order requests, cutting order processing time by 40%
- Developed REST APIs that supported 1,000+ daily transactions, ensuring seamless data sharing and boosting system efficiency by 30%
- Leveraged MongoDB for caching API responses, reducing API response times by 20% and handling high-traffic requests exceeding 5,000 per day
- Automated weekly application deployments using Jenkins Linux scripts, decreasing deployment failures by 90%
Education
Master of Science - Computer Science, Machine Learning and Scientific Computing
San Diego State University
San Diego, CA05.2026
Bachelor of Science - Electrical Engineering
University of Illinois At Urbana Champaign
Champaign, IL05.2021
Skills
- Languages: Java, Python, C, JavaScript, Nodejs
- Frameworks/Tools: React, LangChain, PyTorch, Docker, Jenkins, GitHub, SVN
- Databases: SQL, Oracle, MongoDB, Azure Cosmos DB
- Cloud: AWS (P4D), Azure
- Concepts: Deep Learning, LLMs, Computer Vision, NLP, API Development, Microservices
Timeline
Project Lead - Crowd Flow Simulation
SDSU Research
12.2024 - 05.2025
Applied AI Research Assistant
San Diego State University
08.2024 - Current
Java Full Stack Developer
GUESS Inc.
08.2021 - 07.2024
Master of Science - Computer Science, Machine Learning and Scientific Computing
San Diego State University
Bachelor of Science - Electrical Engineering
University of Illinois At Urbana Champaign
Similar Profiles
Lindsey TortoraLindsey Tortora
Student Services Professional II-Academic Advisor at San Diego State UniversityStudent Services Professional II-Academic Advisor at San Diego State University
Dylan Rafael YamzonDylan Rafael Yamzon
Intern at San Diego State University ResearchIntern at San Diego State University Research
Vince VelascoVince Velasco
Computer Science Student at San Diego State UniversityComputer Science Student at San Diego State University