Sep 2020 – Jan 2022 · 1 yr 4 mos

• Role: Technical Architect/Visionary
• Team Size: 3
• Intention: To create the prototypes/ production level software applications to productize some of the new ideas(algorithms) to a professional product to exhibit the competence of the client
• Tools: Intel OneAPI, DPC++, C++(14), SOLID patterns, Angular, Node JS
• Domain: Software Design, Machine Learning/AI
• The target is to productize one or more novel M.L algorithms to create a fully production version and exhibit the product for business.
• The productization requires an application to be developed from zero with multiple features and requires SOLID design patterns. The aim of this project is to showcasing various products (Machine Programming, Confidential Computing) from various teams of the client.

Apr 2020 – Sep 2020 · 5 mos

• Role: Technical Architect/Visionary
• Team Size: 3
• Intention: To create the prototype to convert speech to speech translation from English to German using OneAPI eco-system programmable with DPC++ (extended from SYCL).
• Tools: Intel OpenVINO, OneAPI, DPC++, Python extension into C++, Translate API library
• Domain: Machine Learning/AI
• The Overall idea is to provide an intelligent language translation from the Video feed, which shall not have translation content provided by the content providers to be used in the embedded systems such a TVs or to implement the intelligent translation for Online Media Services Provider and production companies such as Netflix.
• The first prototype is being implemented which shall translate the content language of audio only, only to the fixed set of hierarchically similar languages without having to keep the integrity of the source language grammar.
• OpenVINO is used to convert audio to text, Translate AI to be used for Text to Text a subsequent speech synthesizer to get the final output.

Dec 2019 – Apr 2020 · 4 mos

• Role: Technical Architect
• Team Size: 5
• Intention: To create a new application cross-platform, in Windows, (also using Windows Skills UWP) and in Linux (Ubuntu) that integrates the Face detection/identification and human tracking and Speech Recognition features with the Host UI (Skills UI and WINFORMS in Windows), QT in Windows and in Linux with a novel customizable and modular cross platform framework (implemented in C++17), using OpenVINO based library
• Tools: Intel OpenVINO Linux OS, Windows, Software Design Patterns, QT and Winforms
• Domain: Machine Learning/AI, Software Architecture Design
• Software architect and technical mentor to productize over the NOVEL framework, to integrate the Face detection/Human Tracking AND Speech recognition features with host UI using Intel OpenVINO
• Designed and developed the basic cross-platform modular, customizable framework in C++17 to implement Commands, which are capable of Executing, Undoing Journalizing and Terminating, Event Handlers which are capable of notifying and observing those events, etc. in a record 2 months with very less input.
• Integrated features that are implemented using OpenVINO and Python libraries with C++.
• Mentored the team to extend the basics to improve the product with new features that could be loosely coupled with the framework.
• Export ready APIS, which could be used by wrapping with JAVA, PYTHON scripts.
• Strong understanding, fundamentals and have yet to be published papers in machine learning and AI, related to CNN, KNN, SVM with Python, C++.
• Study to decouple the computation from client to make the APP cloud enabled. As the first step, study is being to make the computation engine hosted by a server and connected to by clients over web.

Mar 2019 – Dec 2019 · 9 mos

• Role: Technical Architect
• Team Size: 7
• Intention: To implement PIGMS paradigm
• Tools: C, WinDbg, GD2, GIT, JIRA
• Domain: Graphics System Process
• Implemented Process Isolated Graphics Memory Spaces in Windows 7 –
• Good theoretical understanding of device drivers, especially related to system level graphics (KMD-UMD).
• Lead implementation of the Per Process Graphics Translation Table (PPGTT) to isolate the graphics process space from Kernel memory there by isolating the failures of the processes only on to themselves rather on the OS.
• Create and Tracked Project executions via, JIRA, and DevOps Tools.
• Created new applications for analyzing KwWork, static code analysis and graphics sleep/hibernate tests using C#.NET using WINFORMS.

Nov 2017 – Mar 2019 · 1 yr 4 mos

• Role: Technical Architect
• Team Size: 5
• Intention: To enhance the 3D scanning and inspection (Metrology) software architecture for future proofing, to add business competitiveness with full quality with the perspective to unify a suite of CAD-Metrology related products into a single software spectrum framework.
• Tools: C++-11(advanced), C#.NET wrapper for C++, Software design patterns and principles, AGILE
• Domain: Metrology(CAD), Software Architecture Design
• Tasks accomplished:
• Product architecture re-design: Migrated the executable commands to harness unified platform.
• Decoupling UI framework from command execution framework by migrating the entire UI and the workflow to Windows based WPF UI using C#.NET
• Executable Commands made completely parametrizable by implementing C++ interfaces using C++/CLI(C#.NET) to couple with C#.NET WPF UI.
• to support flat file as well as to create modern database components to persist data
• Provision to migrate the current flat file DB to advanced non-RDBMS DB

Jun 2016 – Nov 2017 · 1 yr 5 mos · Bangalore, India

• Role: Software Engineer -Application Development (as Architect)
• Team Size: 10
• Intention: To enhance and support the architecture of Total Airspace and Airport Modeler (TAAM).
• Tools: C++-11, C#.NET wrapper for C++, Software design patterns and principles, AGILE, OpenGL, Microsoft Visio
• Domain: Software Architecture Design, Avionics
• Accomplishments:
• Continually added business competitiveness to the product TAAM with enhancements in the technology, architecture and quality with the perspective from product architecture.
• Analysis and proof of concept to migrate the graphics viewer from OpenGL to more sophisticated rendering engine {Unity 3D and Unreal for POC) by creating new architecture enhancement, by decoupling the existing simulation with rendering engine from the common UI and migrate the WxWidgets based UI to WINFORMS based C#.NET and WPF.
• Analysis done to use TAAM as Software as a Service by re-architecting the product TAAM, to enable the legacy product TAAM to be web enabled.

Feb 2015 – May 2016 · 1 yr 3 mos · Pune Area, India

• Role: Software Engineer -Advanced
• Team Size: 15
• Intention: Feature enhancements and continuous maintenance of NX Part Modeling
• Tools: C++, C#.NET wrapper for C++, Software design patterns and principles, Computational geometry for curves and surfaces, Object Modeler in house architecture, Microsoft Visual Studio
• Domain: CAD – Advanced algorithms
• Accomplishments:
• Fixed some of the very complicated defects whose cause was deep routed mathematical error.
• Part of the team which enhanced NX architecture to re-write Trim curve (3D) functionality
• Developed a new feature (prototype in hackathon) to reconstruct a solid's point cloud data to STL solid format using very optimal memory management algorithms and very efficient point cloud query

Jun 2013 – Mar 2014 · 9 mos · Goa, India

• Accomplishments as Architect:
•  Cross-Platform 3d printer connection architecture (implemented in C++) to enable cross communication between Windows desktop, Phone, Apple, iOS and Android platforms and (any) 3d printer.
•  Created a basic Model Viewer Controller (MVC) Cross-Platform Architecture to give the user basic viewing and editing capabilities of the model to be 3d printed using OpenGL ES across platform.
•  Adapter layer to route the geometry operations from across platforms to the underlying kernel.
• Accomplishments as Feature Developer:
•  Geometry kernel level operations on meshes such as Boolean operations and mesh segregation
•  Creation of Viewer objects in OpenGL ES while using Rhino 3DM memory model as neutral data format for the Model.
•  Lead Developer in implementing the C++ to the native language adapter for Windows (C# .NET), Objective-C for Mac and iOS platforms and Java (JNI) for Android platforms.
• Strong suits: C++, Software Design Patterns and Principles, C++ to native language adapters in Windows Desktop ( C#.NET/Managed C++, Objective-C in Mac/iOS, JNI in Android ), geometric algorithms.

Dec 2010 – Apr 2013 · 2 yrs 4 mos · Bangalore India

• Geometric modeling of the die/punch for sheet metal stamping analysis.
• Worth mentioning major feature enhancements:
• > Multiple die/punch modeling including symmetry
• > Modeling of final stages of die/punch modeling: filleting
• CAD Modeler : Visual Die Maker
• Solver : PAM STAMP 2G
• Technologies: C, C++, VC++, MFC
• Architecture: In house Visual Architecture

Sep 2008 – Oct 2010 · 2 yrs 1 mo · Singapore

• Software Development Engineer,
• Team member in the AutoCAD PSEB-Engineering development team.
• 1. Geometry feature enhancement and stabilize two releases of AutoCAD
• 2. Part of the team which started realize Design Computation ( Generative Shape Design) in AutoCAD.
• Tools: C++, C# .NET/Managed C++, ObjectARX

May 2007 – May 2008 · 1 yr · Pune/Pimpri-Chinchwad Area

• Team member in Research and Prototype projects in 1. Rule Based Configuration. 2. Integrated Functional Layouts, and 3. Silhoutte curve computation
• This work involved implementing and geometrical algorithms
• Tools: C++, C#.NET, OpenGL,ObjectARX(AutoCAD plugin tool)

Sep 2003 – Apr 2007 · 3 yrs 7 mos · Bangalore, India

• I had been working on many products.
• 1. IGRIP ( Interactive Graphic Robot Instruction Program ), a device, mechanism and robot modeler to model processes. This is entirely written in C, a marvel product which is very stable. This is owned by Dassault Systemes.
• 2. Device Building and Workcell Setup Workbench of Delmia, a process modeling tool of Dassault Systemes. Worked in Device and Robot Building, uses CAA ( Component Application Architecture ) based on COM and CORBA programmed in C++.
• (a) . Worked in Inverse Kinematics of industrial robots,
• (b) Computational Geometry operations on tessellated data, such as Meshing, Boolean Operations, collision detection and obstacle avoidance
• (c) Modeled robots to have multiple chains rather single chain. This is the kind of work I liked and enjoyed. This involves enough knowledge of position kinematics, inverse kinematics, World Data Model simulation architecture of Delmia, Transformations in CAD basic geometry operations.

Feb 2002 – Sep 2003 · 1 yr 7 mos

• I had been working on
• 1. path planning( forward and inverse kinematics, collision detection and obstacle avoidance ) and
• 2. synthesis of snake like robots ( Hyper redundant manipulators ). This involved numerical optimization of multiple objectives. Synthesis involved the problem of solving by combinatorics for 2D.