Across industries as different as industrial wireless networks, cybersecurity, environmental monitoring, and agriculture, one challenge keeps repeating itself: enormous volumes of raw data exist, but very little of it becomes a decision anyone can act on in time. Nishi Tadamalla has spent the last several years building the bridge between the two. A software developer and data analyst by training, with a Master of Science in Information Systems from Wilmington University, Tadamalla has quietly assembled a body of independent research that tackles this problem from four unrelated angles, and each project reveals the same instinct: find the bottleneck between data collection and decision-making, and design a system whose real-world impact is measured not in theory but in devices supported, threats detected, or dollars saved.
The most technically ambitious of these projects addresses a problem that sits at the intersection of two fields that rarely speak to each other, wireless communications and control theory. In “Control-Aware Low-Latency Scheduling for Time-Sensitive Wireless Networks,” Tadamalla proposes a scheduling method, referred to as CALLS, that rethinks how industrial control systems communicate over Wi-Fi. Modern factories and robotics platforms increasingly rely on the IEEE 802.11ax wireless standard to connect sensors and actuators, but conventional scheduling treats every packet the same way, regardless of how urgently a control loop actually needs it. CALLS instead ties the wireless scheduling decision directly to the stability requirements of the control system itself, using Lyapunov stability metrics to dynamically adjust packet delivery targets and allocate resource units only where they are needed most. Tested against simulated control systems such as inverted pendulums and balancing boards, the approach allowed roughly twice as many devices to operate reliably on the same network compared with conventional scheduling. That doubling in capacity is not a marginal gain; it is the difference between a factory floor that must ration its wireless connections and one that can scale sensor and robotics deployments without re-engineering its infrastructure.
A second project moves from industrial infrastructure into cybersecurity, an area where visibility itself is often the scarce resource. Peer-to-peer botnets, unlike the centralized command-and-control networks that defined earlier generations of cyberattacks, distribute instructions across thousands of independent nodes, making them resilient to the simplest countermeasure of shutting down a single server. In “Real-Time Visualization and Optimization of Peer-to-Peer Botnets for Efficient Management and Propagation Control,” Tadamalla builds a controlled, fifteen-node P2P botnet environment and develops a real-time visualization layer that tracks network topology, node status, and route efficiency as the network evolves. The practical payoff is speed: security teams no longer have to reconstruct an attack after the fact, but can watch commands propagate live and intervene before a network reaches full strength, a capability directly relevant to defending against the kind of large-scale coordinated attacks that botnets like Mirai and Kraken have historically enabled.
A third strand of research turns the same analytical instinct toward environmental monitoring. In “Evaluation of Approaches for Identifying Mining Activity through Hyperspectral Imagery and Geographic Information Systems,” Tadamalla evaluates the software tools available for processing AVIRIS hyperspectral imaging data, comparing Spectral Python, HyperSpy, and the MATLAB Hyperspectral Toolbox before settling on Spectral Python for its neural network compatibility and active maintenance. The paper goes on to compare mineral classification, GIS-based comparison, and a hybrid
of the two, concluding that mineral classification alone offers the best balance of accuracy and computational efficiency for large-scale, real-time monitoring. By replacing slow, expensive manual field surveys with an automated framework, the work opens a path toward an open-access tool that regulators and environmental researchers could use to track deforestation, habitat loss, and water pollution from mining as it happens, rather than months later, directly strengthening real-time environmental oversight and regulatory compliance.
The fourth project trades satellites and sensors for spreadsheets, applying the same data-driven rigor to an industry not usually associated with machine learning: Australian wine growing. In “Decoding the Economic Forces of Australian Vineyards,” Tadamalla analyzes a ten-year dataset of over six thousand vineyard records, collected through the Sustainable Winegrowing Australia program, using the XGBoost algorithm to isolate the factors that most strongly drive operating costs and revenue. The analysis, visualized through Sankey and chord diagrams, identifies water and fuel consumption as outsized cost drivers and surfaces significant regional variation, at a moment when the industry has been squeezed by pandemic-era labor shortages, freight disruptions, and a nineteen percent single-year export decline tied to trade tariffs. By translating those findings into concrete, region-specific benchmarks, the research gives individual growers a way to cut costs and improve yields using data they already collect, turning sustainability from a compliance obligation into a measurable competitive advantage.
Taken together, these four projects do not share a single industry, but they share a method and a measure of success: identify where a system’s complexity outpaces a human’s ability to monitor it in real time, and design the analytical layer that closes that gap in a way that shows up as faster response times, wider device support, stronger environmental oversight, or a healthier bottom line. That pattern, repeated across wireless networks, cybersecurity, remote sensing, and agricultural economics, is what distinguishes Tadamalla’s body of work from a conventional data analytics résumé, and points toward a research trajectory built on solving the same underlying problem, and delivering the same tangible impact, wherever it appears.
