Ranjan Rishi Chambial

• Research & Development for Distributed Computing Systems, related to AI & Deep Learning Frameworks on Heterogeneous Computing Platform.
• Distributed AI/Big data: movement and management of cloud computing frameworks such as Spark (utilising Java and Scala at the core), Hadoop, Ray, Neural Network Model Sharding using tensorflow & Pytorch for distributed sytems, Multi Node Model training and inference(via GRPC protocols).
• Code Base Diagnostic, tracing for various datastructures, Benchmark and Profiling tool development (C++ & Python), Dask library for data handling and post processing, data analysis using Pandas, Scikit learn.
• Datastructure performance and optimisation based on various combinations of AVX instruction set and compiler optimization techniques. Analysis of the Assembly and memory dump of the instructions.
• Architectures used: ARM(Cortex), INTEL Xeon Phi, CUDA compliant GPUs.
• Profiling the code for benchmarking for Power and performance on heterogenous computing platforms(CPU and GPU)

• Designing systems, processes, methodologies and vehicle designs to enhance overall vehicle performance for the customer. Transforming concepts into prototypes for testing, validating and improvement for mass volume production.
• Projects were done for following clients: Toyota, BMW and Daimler

Key Projects:

• Radar Module with AI functionality for ADAS, Radar Calibration and developing algorithms to perform inference on the data collection via Radar point clouds and signals
• Google Test, Git, Clustering algorithms, UML Generation for codebase, Performed static code analysis.
• DSP algorithms and code optimization as per Processor architecture.
• Incorporation of Radar Meta Data and Inference output signals with ECUs and Drive train decisions; Testing the Radar functionality & developing testing framework using Behave Python framework.
• Deployed the tests using docker(dockerfiles) and bitbucket pipelines comprising the tests along with the setup sequential routines on the testing device.

• Build and implement state-of-the-art computer vision algorithms and software by understanding research papers. Research and execute sensor fusion/perception algorithms (radar, LIDAR, and camera), motion planning, and control and estimation algorithms (GPS, IMU’s)for the autonomous drone models. Develop Deep learning based solutions for both drones and static industrial cameras of general inference.
• Projects were done for following clients: Shell, Bharat Petroleum, Adani Power Ltd etc

Key Projects:

• Project: Detecting, Tracking and Counting People’s movement in the area of Interest- The project was built on Gstream framework and Nvidia’s deep stream SDK 4.0. The application was designed in C, C++, TensorRT & CUDA.
• Project: Tracker Comparisons- Testing the robustness of various trackers against MOT challenge 2016 videos. Comparison was made by evaluating the components like Predictor(Kalman filter), Association algorithms of each tracker.
• Project: Neural Network End to End Training - Used Tensor flow & Keras. The VMs & GCP Instances were used to train large neural networks on cloud.

Network 1: Corrosion/Rust/Paint Damage Detection & Classification.

Network 2: FaceMask detection & classification(During COVID crisis)

• Project: Flare and Smoke Detection - Conventional CV techniques background subtraction & GMMs fitting.
• Added additional functionality on top of the tf-explain package used for neural network interpretability - Added GRAD CAM PLUS Visualization functionality that can be added as a tensor flow call back while training the network.
• Developed Tools/Scripts for use cases such as: Multi class CAM visualization; Training Data preparation (Augmentation, Synthetic Data Preparation, Web Scraping, Patching of images)

• Coordinated a wide range of Microsoft’s client projects from inception to completion, programming done using C/C++/C#, visual studio, Windows Internals, and .Net framework. Handled various chandelling cases involving CLR, NGEN, and JIT compilation.
• Assisted various clients in fixing technical glitches related with C/C++ & STL usage, Identified user difficulties of clients using C#/.NET framework such as managed memory concepts & Garbage collection (Dangling pointers/roots).
• Managed memory management issues related to GC, empowering developers with the knowledge of Windows Internals such as Virtual Memory used by OS MMU.
• Resolved Kernel Level programming issues(Thread Synchronization)involving multiple threads & data sharing
• Collected memory dumps for faulty application using Procdump, TLS and & troubleshooting tools.
• Worked on WinDbg, Debugging & Visual Studio mixed mode debugging & provided code debugging support to the application developers.
• Handled cases involving CLR, NGEN, JIT compilation and provided first-line technical support to developers.
• Handled cases involving Microsoft native technologies, such as MFC, COM programs & Winforms. Responsible to improve the efficiency & performance of software code such as runtime and UI responsiveness.

• Led the design, development, and installation of PCs and workstations by utilizing software diagnostic tools, validated software tools on target systems,henceforth preparing Release Documentation.

Key Projects:

• Developed software for System Diagnosis, both for Servers & PC’s running on Windows OS, structuring program for developing a software tool and scaling up.
• Recording & evaluation of additional user requirements (using Requirements Engineering), Implementing software test modules communicating directly with onboard hardware modules.
• Software diagnostic tools covered components such as core, cache, memory, bus architectures, IOs etc.
• Power monitoring for series of onboard CPUs on servers.

• Worked with Multi touch Diffused Illumination technology.
• Was part of innovative products consisting end to end Multi touch, Gesture, 3D projection, Home automation Projection Mapping, Advanced display & Hologram solutions.
• Worked with MultiTaction-one of the partners of Techrobotix in Gitex Technology week & learnt their Cornerstone SDK.

• Worked on a Rugged computing solution for Dubai Electricity & Water Authority, installation/deployment & site visits along with the team.
• Site visits to onshore oil rigging companies (Sharjah Area) & learnt about the implementation of Rugged computing.
• Grasped a brief knowledge about the Rugged computing hardware & various thresholds of reliability tests. Researched on various embedded computing standards used in this field such as various ICs that sits on top of rugged computers & different cooling systems to help the computers cope up with the high/extreme climate conditions.