Building pipelines to transform, analyse, visualize and maintain A320 Family aircraft production data. Create ML solutions to support several applications in and around the assembly process. Develop Javascript apps for internal tools

Behaviour logic algorithms of aircraft in combat, Combat mission planning (GOAP, HTN), resource allocation problem, developing models for simulating sensor (Radar, Flir etc.), missile models, guidance laws and further functionalities involved in air to air combat. (Python and C)Developing robust swarming algorithms from mission execution perspective, considering dynamics of remote carriers, arrival times, signature optimal flying, collision avoidance, split-up and recombination, receding horizon.Developing reinforcement learning algorithms, cost functions for AI pilot, for learning new air to air combat maneuvers. Putting efforts towards fast converging algorithms for rapid testing of the cost functions (Tensorflow, Keras, Python)

Working on developing stable, numerically inexpensive swarming algorithms for UAVs to perform tactical missions. Centralized/ Decentralized path planning with collision/ obstacle avoidance and formation control for Multi-Agent systems.

Developing an efficient Weather Avoidance algorithm for a Remotely Piloted aircraft.Developing fast and efficient path planning algorithms considering the rules of air and airspace for Upper and Lower airspaces. Improving previously developed algorithms to incorporate dynamic weather conditions like Thunderstorms and remain predictable in order to reduce airspace violations during maneuvers. Algorithms consider cases with availability of the whole weather data as well as line of sight weather data (RADAR)The algorithms include:- Specially modified A* algorithm with dynamic properties- Combination of RRT and A* (I call it RRT-A*) considering the best of both the algorithms- Standard Non-Linear Programming with an innovative system model which makes it easier to develop a special solver from scratch- Implementation of Deep Reinforcement learning (Q Learning) with an intelligent reward function in order to find the optimal path

Develop and model control strategies for components of HVAC like room temperature control, boiler etc. in Dymola for energy efficient buildings. Implementing several models for HVAC components like Cooling Towers, Mixing Circuits, Non-Mixing circuits etc. and testing these models for expected behavior.

Developing a Simulink based Hydraulic Tool Box with basic elements to build a complex hydraulic system.

Economic and performance aspects of blended coal in a typical coal fired boiler