Research Assistant

• Developed learning-based probabilistic robotics methods for reliable navigation under partial observability and cooperative multi-robot localization
• Designed SCoDA, a shielded conditional diffusion framework that learns task-specific fiducial layouts from closed-loop rollouts and steers generated placements toward regions where localization corrections improve trajectory execution
• Built procedural simulation and hardware-evaluation pipelines for randomized environments, stochastic disturbances, trajectory planning, and physical fiducial deployment
• Achieved 90% task success and 95% waypoint following using the minimal fiducial budget, with layout generation in under 1 second versus hundreds of seconds for rollout-based search
• Developing decentralized belief estimation using neural belief propagation and normalizing flows to fuse landmark, inter-robot, and teammate-message observations

Industry-Sponsored Research

• Built MAESTRO, a multi-agent assistant translating supply-chain optimization outputs into transparent, source-grounded explanations for non-technical stakeholders
• Developed a 5-stage LangGraph agentic-RAG pipeline coordinating 4 role-based expert agents through retrieval, tool use, validation, and answer aggregation
• Integrated Pandas and Excel analysis tools to extract KPIs, throughput summaries, and other quantitative facts from solver spreadsheets
• Implemented 3 system configurations: mixture-of-experts, traditional RAG, and agentic RAG
• Open-sourced the system with stakeholder-facing reports and evaluation artifacts (https://github.com/Manavvora/MAESTRO)

Research Assistant

• Evaluated machine-learning methods for multi-object interstellar trajectory design using synthetic trajectory data and mathematical guarantees to support mission-target prioritization
• Utilized public astrophysical databases to prepare stellar data for AstroNN-based spectrophotometric-distance estimation
• Evaluated how distance estimates could support interstellar target selection and trajectory planning

Data Engineer Intern

• Automated data-quality checks with Apps Script, enforcing shared standards and saving 30-60 minutes per case
• Developed a flexible app-version reporting program consolidating analytics from multiple applications, environments, services, and report types into a single email
• Automated report-content management, improving usability and saving engineering time spent maintaining reports manually
• Utilized SQL to prepare, organize, and convert raw data into usable information for data analysts

Flight Operations Controller Intern

• Planned aircraft movements within a three-day operational window addressing unforeseen challenges such as aircraft damage and high-season demand
• Optimized schedules to accommodate additional flights with limited aircraft availability
• Coordinated with pilots, air traffic control, refueling teams, and other departments during takeoff and landing to ensure seamless operations
• Managed real-time communications for schedule adjustments including early/delayed departures and facilitated emergency procedures such as rerouting to alternate airports

• LGBTQ: NA
• Ethnicity: NA
• Disability: NA
• Veteran: NA
• First-generation College Student: NA
• Pell Grant Eligibility: NA
• Work Authorization: NA
• Visa Sponsorship: NA