Nitin Singh Rathore — Software Engineer, Arlington TX

Arlington, Texas

• Support computer science coursework through lab guidance, student mentoring, grading, and technical explanations.
• Help students strengthen programming, debugging, data-structure, and systems fundamentals through practical examples.
• Continue research and project work around reliable coordination for autonomous UAV networks.

Indore, India

• Architected and maintained 6 Java/Apex backend services using object-oriented design principles and modular service boundaries, powering structured business workflows across 3 enterprise clients and reducing manual processing overhead by 40%.
• Engineered REST and SOAP API integrations between 5+ internal and external systems, achieving 99.8% uptime across production integrations.
• Optimized 15+ SOQL/SQL queries through indexing strategy and access-pattern redesign, cutting high-latency production query times by 35% and improving overall API response time.
• Authored 500+ automated and manual test cases with JUnit, Postman, and Tosca, slashing post-release defects by 30% and reducing QA cycle time by 2 days per sprint.
• Resolved 12 critical production incidents through systematic log analysis and root-cause debugging within 2-hour SLA windows.

Indore, India

• Accelerated module-level processing efficiency by 15% by refactoring 4 core Java/SQL data-access modules, reducing average query execution time from about 320ms to 275ms.
• Eliminated 20+ pre-production bugs by authoring JUnit-based unit and integration test suites, preventing regression defects across 3 release cycles.
• Shipped 4 backend features into production through Git-based PR workflows and structured release processes, contributing to 2 major quarterly releases.

A full-stack career search system built around fast document indexing, ranking, and background processing.

• Indexed more than 31,200 documents for search and ranking workflows.
• Improved median lookup time from about 690ms to 3.5ms at a 32k record scale.
• Used Redis/RQ workers and Dockerized services to keep ingestion, ranking, and API work reliable.

A real-time AI meeting and workflow assistant that turns live conversation into searchable notes and follow-up actions.

• Built meeting transcription, summary, and retrieval flows with Deepgram, LLM APIs, and pgvector.
• Reached about 94% NLP/RAG precision with p95 response latency under 3s.
• Connected GitHub and Google APIs for workflow automation with about 98% successful task creation and no duplicate actions.

A distributed web-scraping queue for high-throughput, idempotent task execution across cloud workers.

• Processed about 1.4k tasks per second with Redis-backed queues and containerized workers.
• Added idempotency and recovery paths that restored workers in under 10s with zero duplicate task execution.
• Provisioned repeatable AWS infrastructure with Docker, Kubernetes, and Terraform.

An autonomy assurance framework that estimates localization trust and adapts UAV planning, supervision, and offboard control before unreliable state propagates through the stack.

• Built continuous localization trust estimation with authority modes for nominal, limited-operation, and degraded-hold behavior.
• Integrated trust-aware mission execution with BehaviorTree.CPP, Nav2 planning, PX4 SITL, and Gazebo simulation.
• Added dataset replay plus automated CI, unit tests, and ROS 2 integration tests for repeatable validation.

Research work focused on dependable autonomy for multi-agent UAV systems, including coordination, trust-aware supervision, and validation under degraded sensing or communication.

• Build simulation and validation workflows for UAV coordination, trajectory tracking, and autonomy assurance.
• Explore trust-aware control ideas where agents adapt behavior before unreliable state estimates propagate through the autonomy stack.
• Connect research prototypes with practical engineering through ROS 2, PX4 SITL, Gazebo, and repeatable test pipelines.

M.S. Computer Science Thesis · Computer Science and Engineering · UT Arlington

A cross-layer supervisory control framework for low-altitude UAV swarms operating under dynamic and uncertain conditions.

• Augments a classical PID controller with a lightweight supervisory layer that monitors dynamics, observation, and networking behavior.
• Identifies likely fault conditions and applies bounded adaptations to reference motion, formation geometry, and coordination parameters.
• Evaluated in multi-UAV simulation with controlled fault injection and multi-seed experiments, improving robustness under disturbance and communication faults while preserving comparable nominal performance.

Advisors: Dr. Md Salik Parwez, Dr. Debashri Roy, and Dr. Nadra Guizani.

M.S. Computer Science · Arlington, Texas

Artificial Intelligence · Machine Learning · Cloud Computing · Neural Networks · Software Testing · BATs

B.Tech. Computer Science · India

Data Structures & Algorithms · Object-Oriented Programming · Operating Systems · Database Management Systems · Computer Networks · Software Engineering · Discrete Mathematics · Computer Organization & Architecture