Aakash Khatri

Sole architect/developer of a Python/Django (DRF, Channels) data + support platform used daily by 20+ agents; ~4k–5k Amazon DE tickets/month (~40% volume). Built modular ingestion/mapping and integrated Amazon LWA (OAuth2). Event-driven async processing (RabbitMQ, Celery). Reporting-ready KPI datasets/endpoints (Power BI). CI/CD via GitHub Actions; Azure/Docker. Continuous usage since rollout with no business-interrupting outages reported.

Researched and implemented ML/SE approaches in dynamic fuzzing. Built a Python framework and data-processing pipelines to integrate, benchmark, and analyze new algorithms. Used parallel programming and GPU acceleration, achieving up to 7× performance improvements. Worked across Python, Kotlin/Java, and C++ and supported reproducible experimentation (test harnesses/automation; CI where applicable).

Assisted with research development tasks; built and integrated C++ components and tooling.

Developed object detection (CNNs) and tracking for drone-based inspection and intrusion detection. Built a real-time C++ pipeline using parallel programming and GPU acceleration (~30 fps). Implemented data pipelines for training/evaluation and integrated components into production workflows. Deployed to embedded Linux targets (e.g., NVIDIA Jetson family) and integrated with surrounding systems as needed.

Delivered Salesforce solutions by translating business requirements into scalable implementations. Built backend logic and integrations using Apex and platform tooling; contributed to modular customizations and performance improvements. Worked in Agile delivery environments using Jira/Confluence; supported quality via debugging, optimization, and testing.

- Data Science Suite: implemented de-duplication techniques for web-crawling data in Java. - TH E SI S : Development of a Reinforcement Learning Interface for Partial Differential Equation Solvers using Python and OpenAI Gym.