Manish Sharma

Myself Manish Marsarma, and I'm basically from Mirzapur, Uttar Pradesh. And I have come I completed BCA from Bao Banerjee Das University in 2020, and college based university based and, left now. And after that, I started my career as a junior Python developer in TCS in 2021, Feb February, and I have done around 3 projects over there, and, I, like, I used, like, for, all 3 projects, I used technology like Python, Django, Tornado, and Python, Django, Tornado, and PostgreSQL, HTML, CSS, AngularJS, And yeah. I think the and in September 2022, I left TCS. And after that, I, switched on another company named PwC, where PwC is client company, and my payroll company is MobiFLY. And here, I'm working on and project like, on a project called invoice navigate, where product is generating to, Generating Ewa builds and IRNs. Okay. And from October 2022, I am here. And, like, technology used in this project, Python, AWS, Lambda function component, like, serverless, all component like s three, SCS, SQS, CloudWatch, etcetera, and, Jenkins for CICD, development and, HTML CSS, Falcon for APIs. And okay. That's a post list for database. And after, like yes. that's all about me. Thanks for asking. And, Lastly, my skill sets are, like, Python, Django, Tornado, and, PostgreSQL, MongoDB and, MongoDB and HTML, CSS, Pandas, NumPy, Django ORM, AWS Lambda function, CloudWatch, s 3, and, CloudWatch, as I told you.

In what In what kind? Like, what ways? Reinforcement learning is in a, like, simulation framework like allows, robots to, robots to learn and refine behaviors in a, in a Controlled environment before real world deployment. And this helps, improve outcomes By enabling, efficient training, reducing risks, and optimizing, performance through creative Simulations.

How do I like, Okay. to reduce computation, a load of high Fidelity simulation. like, I have used Many things, like, parallel parallelizations, vectorizations, Caching and algorithm like, algorithmic, improvements, GPU accelerations, Like, utilize libraries such as, Pandas, NumPy, like, and to improve optimize memory, Using, like, memory management, to optimize memory users by minimum. Minimizing unnecessary data, duplicate duplication, releasing unused object, and, efficiency, mani, managing data structures. And

Let How would you optimize a physical Physics based simulation, you know, is Sim to run engine. Right. optimizing, Isaac sim, for real time, performance and, like, performance and high fidelity by, implementing adaptive, level of detail, but, level, like, adaptive level of, detail, parallel processing, GPU acceleration, real time constraints, and optimized, physics, Slower, streaming assets. As in as in the simulation, guessing reduce preset like precision. Tuning simulation parameters and utilized code profiling For, bottleneck identification.

Automated test cases for simulation in Isaac sim using okay. Automate robotics, simulation test cases in Isaac sim using the Isaac SDK Python, like, have, Python have NPL, like, a pytest for the test organizations. Organization scripted scenarios for different or test scenarios, assertion for validation, handle made The handler sorry. Headless, mode for if, Headless mode for efficiency per, parameterizations for diverse testing. CI integration for continuous testing, logging for, information capture and error handling.

For a distributed simulation over multiple GPUs and Omniverse, how would you manage Like, ensure, state consistency in a distributed omniverse, simulation with the synchronization points, message passing, global estate management, time synchronization, checkpoints, mistake, simulation, and conflict resolution, latency,

In context, The provided, code, logic, like, pseudo code has a, Logic error in the void loop condition that may prevent the, robotics arm from reaching the target position. The issue lies in using, like, not equals to operator to check if current position is equal to target. Due to potential, floating points precision issues, it's generally not, reliable to use Exact equality from comparing floating points values. To address this, you should check Whether the distance between current position and target position is below a certain tolerance rather than Relying 1, exact equality. Like, we can, use, We can use while as in a distance, in, small bracket current position, comma, target position, greater than tolerance, and greater than tolerance. Yes. Then I think this will, This will be better for, like

Navigate through a series and explain the bug that Will cause the effort board to skip every way points. The bug is the provided pseudo code that, causes the robot to skip every other waypoints. It's, due to the incorrect uses of a continuous statement inside the loop. The continuous statement causes the loop to, skip the Remaining code in the loop body and move to the next iteration. In this case, it will cause the, loop to skip the weapons at index I.

How would you scale a real time simulation system to manage increase complexity due to additional new robot model. Like, the 2 scale real, time simulation system With new robot models adopt a modular architecture for easy integration, utilize parallelization to distribute, Computational load, manage level of details l LOD dynamically, and introduce a synchronous Processing for noncritical task, optimize, physics simulation for each robot model, implement Configuration parameter for simulation settings use dynamic resource, allocation and Load balancing, monitor resource users and scale resource accordingly. Profile and profile and pro optimize code for performance, conduct, incremental testing and validation for each added, robot models.