Mechanical Design in Robotics with Sara Wojciechowski

 
WomenPodcast_Sara.png
 

Building robots that can interact with the physical world involves a team of scientists, engineers, academics, and business operators. Edaena Salinas, host of The Women in Tech Show, interviews Sara Wojciechowski, a mechanical design engineer at Kindred. Sara shares in this podcast the types of problems that robots can tackle and their current limitations. They talk about the components of Mechanical Design and how it’s applied in robotics. Sara explains how ideas are prototyped and developed for building robots.

Listen to the podcast here

What It's Like Interning at an AI and Robotics Startup

Meet two of our interns, Sachit Ramjee and Ryan Dick. Studying at the University of Toronto’s Master of Science in Applied Computing (MScAC) graduate program, they recently completed an 8 month internship with Kindred in our Toronto location. Here they share the highlights of their internship. Both received and accepted full-time offers with Kindred.


 
Sachit Ramjee (L) and Ryan Dick expect to graduate in 2019 from The University of Toronto’s MScAC program.

Sachit Ramjee (L) and Ryan Dick expect to graduate in 2019 from The University of Toronto’s MScAC program.

 

Q: What did you study in the MScAC graduate program?

Ryan: My studies have been broadly focused around machine learning. In particular, I am interested in the application of machine learning algorithms to computer vision problems. My internship at Kindred was an excellent opportunity for me to do applied research in my area of interest.

Sachit: My studies focused on systems engineering. Specifically, I looked at the design and implementation of different types of computer systems and ways to improve their dependability.

Q: How did you hear about Kindred?

Ryan: I first heard about Kindred through conversations with other students at the University of Toronto. Kindred has built a reputation in Toronto for being a company on the forefront of machine learning and robotics research.

Sachit: I learned about Kindred from fellow colleagues. Hearing about all of the interesting projects that Kindred was working on made the internship sound like a very exciting opportunity.

Q: What was your internship project focused on?

Ryan: My internship project was focused on continuous processing of video media for robotic control. The typical convolutional neural network (CNN) models that have historically been used in Kindred's robotic systems are excellent at extracting information from individual images, but are sub-optimal for interpreting the additional information that is encoded in video streams. During my internship, I explored approaches to real-time video processing with the goal of improving the robot's ability to understand its environment and make decisions.


Sachit: My internship project focused on improving the testability of robotics software. In particular, automated end-to-end testing is a challenge due to the software's dependence on physical hardware and human interaction. As a result, manual testing is needed which can be quite labor intensive and full-system tests cannot be run frequently, as part of continuous integration, for example. The goal of the project was therefore to develop a framework that would enable automated end-to-end testing with or without the physical hardware and human interaction.



Q: What is a favorite moment during your internship at Kindred?

Ryan: It's not easy to pick a single favourite moment, but I'd have to say that it was the Applied Research in Action event hosted at the University of Toronto. This event was the culmination of my internship project, where I presented my work in the form of a research poster. Most of the Kindred team attended the event to support me, and to check out some of the other exciting applied research being presented at the event. It was fun to show off some of the cutting-edge work that is happening at Kindred, while also having the chance to see what my fellow UoT students had achieved through their Master's research.

A close second place: the weekly reading club meetings organized by the Machine Learning Team. Every week I looked forward to learning about a recent paper and discussing it over lunch with the rest of the team.


Sachit: My favorite moment during the internship was being able to see the project come together at the end, especially after spending a few months working on it. Also, being able to work with some great people along the way was very enjoyable.



Q: Which Kindred value do you identify with the most and why?

Ryan: Be Humble is the Kindred value that resonates with me most. Everyone at Kindred makes an effort to express praise and gratitude whenever someone does something that goes above and beyond. This value helps to maintain a tight team-oriented culture at Kindred.


Sachit: I identify most with the value Courage to Innovate. This value encourages taking risks and seeing different perspectives which ultimately helps find new and innovative ways to solve a problem.



Q: What advice would you give students who are interested in working at Kindred?

Ryan: I worked with a team that was very supportive, but, at the same time, no one is going to hold your hand. My advice is to take the initiative to come up with new ideas and drive your own projects. Kindred provides the tools and support necessary to be successful, but in the end, the outcome depends on you.


Sachit: Kindred is a fast paced startup with many unique challenges. Therefore, being able to adapt and pick up new skills quickly is important.



What It's Like Working with Robots Everyday

OnFloorAssociate_Blog.png

Have you ever wondered what it’s like to work side by side with robots every day? We recently sat down with Chere McGreger-Ketchum, an on-site operator who works with Kindred’s SORT robots on a daily basis, to find out what her experience has been like.

Q: Chere, how long have you worked in retail?

A: I have worked in retail for 10 months now. Working with Kindred’s SORT is my first retail job. Prior to this, I was a bill collector, a dental assistant and a manager for a fast food chain restaurant. I also have a cosmetology license.

Q: What's the first thing you noticed about SORT?

A: How unique SORT looks to operate, which many people are intimidated by.

Chere standing next to a Kindred SORT robot

Chere standing next to a Kindred SORT robot


Q: What was the thing you found most intimidating about learning how to work with SORT?

A: I would have to say the most intimidating thing with SORT would be keeping up with it! I make it a challenge every day to see how much harder I can push myself to keep up with SORT.


Q: What is your favorite part of working with SORT?

A: I would have to say my favorite part is the daily challenges it brings. I have had to help assist in fixing a couple things on SORT so it can be in the best working state possible. There is never a dull moment. All the new things I have learned working with SORT, I truly have to say this is my favorite job ever.


Q: What is the best part of working with Kindred?

A: As an on-site operator, I feel I am the eyes, ears, and arms for SORT when the pilots need it. I really enjoy working with the Kindred team because everyone working together makes everything so much easier. I also get to train new on-site operators at other customer sites. It’s such a great experience teaching someone new how to work with the SORT machines. It was so cool explaining everything that I have learned so far. In addition to passing such a great experience onto someone new, it makes it all the better that the person I was training is still with Kindred and she loves it!


Q: What new skills have you learned while working with SORT?

A: I have learned mechanical skills including how to change a gripper, how to reset cameras for the pilots, how to replace cameras and barcode scanners, but I have to say my favorite thing is when I need to work with the electrical panel. I have always loved learning new things and have to say I have learned so much working with SORT. Never a dull moment in the life of SORT.

The Importance of Hiring for Our Company Values

Values_JustineBlog.png


At Kindred, our values are the foundation of our company culture and priorities.  Our core values drive our daily operations, provide guidance on how we communicate with internal and external partners, and serve as a reminder that every decision we make aligns with our company’s mission -- paving the way for a world filled with more powerful and practical AI systems.

Since 2014, we have been developing AI systems that can operate in a complex environment.  After several iterations and pivots along the way, our team never lost focus of bringing to market a product that would make a real impact on an industry.  Our intelligent robotic system, Kindred SORT, operating in production in multiple customer fulfillment centers, is a direct result of our team’s tenacity and practice of our value: COURAGE TO INNOVATE.  

We have also been mindful about the ways in which we integrate our core values with our day-to-day experience.  We created an internal channel in Slack called #Praise that gives our team the opportunity to practice the values of BEING HUMBLE and PLAY & CELEBRATE through the expression of gratitude and appreciation of others.  Cultivating a workplace that promotes expressions of thanks and positive communication, encourages our team to show appreciation for one another and ultimately, helps drive our business goals forward.

We believe in creating a work environment that fosters creativity and new ways of thinking and are looking for additional team members who want to help promote and develop our collaborative work environment.  If you’re interested in drawing from your unique experiences and enjoy encouraging others to do the same - we’d love to hear from you!

Take a look at our careers page for our open positions in our San Francisco and Toronto offices.  If you don’t see an open role that’s a perfect fit, send your resume to our Hiring Team at jobs@kindred.ai

Your Time is Now: Prepare for the Next Peak Season

Holiday_Jim.png

As retailers battle through the remainder of this year’s holiday peak season, there’s one question I consistently hear across the board from all retailers: How do we stay competitive and deliver a delightful customer experience that keeps them coming back again and again? After having experienced 13 holiday seasons throughout my career, the answer lies in speed and accuracy.


So how can you prepare for the next peak season? First step – start now! Second, AI powered robotics are a must if you want to remain relevant. Here are three things to start now to prepare for the next peak:


1. Reduce the Burden of the Staffing Shortage  

AI powered robotics make your processes more efficient and require less manpower in your distribution centers. One robot has the potential to fill several employee shifts, allowing you to level-up your workers’ skills and hire less people during peak season.


As an example, our  customer Gap, Inc., uses robots to sort items at its distribution center in Gallatin, Tenn. In previous years, the company had to increase its human workforce by six times to keep up with demand during peak seasons. After putting robotics in place, they now require much less workforce ramp up during peak periods.  

2. Fulfill Orders More Quickly

End-consumer happiness is a KPI for online retailers and it starts on the warehouse floor. With some major retailers offering same-day or two-day shipping options, the pressure is on to fulfill orders quickly.


The reality is that humans get sick, they need breaks, they can’t work 24 hours straight and unfortunately, they quit. The beauty of robots is that they constantly work at a consistent pace without a shift change, distraction, taking breaks or falling ill, which helps you meet your throughput goals. Robots provide consistent output, and in turn, faster order fulfillment. That can be the difference between a bad review and a happy customer.

3. Increase Order Accuracy

According to one survey of 300 supply chain and distribution managers, fulfillment centers across the U.S. lose an average of $585,000 per year due to mis-picks. While humans are the most accurate workers, mistakes happen when employees are assigned to mundane, repetitive tasks such as picking. Robots with AI combined with automated scanning capabilities guarantee that product data and order fulfillment is accurate.

Now is the time to prepare for the next peak season to delight your customers to drive repeat purchases. Using intelligent robotic technologies, you can grow your operations as demand increases, create a dynamic workforce that marries automation with human skills, and increase throughput and accuracy.  We’re here to help!

Interview with Anurag Maunder, SVP of Engineering at Kindred.ai

Check out our latest interview with our SVP of Engineering, Anurag Maunder, where he shares advice on how to navigate a career in engineering.

 
Location:  San Francisco, CA  Current Role:  SVP of Engineering at Kindred.ai

Location: San Francisco, CA
Current Role: SVP of Engineering at Kindred.ai

 

MANAGERS CLUB (NOV 2018)

What’s your background and how did you get into management?

I did my Ph.D. in Computer Networking and joined Bell Labs. After a few years, I realized that I have a lot of passion for building products. This took me to Bay Area. Becoming a manager was quite organic. I proposed a project and was soon leading a team of engineers as a Tech Lead. I did this for a couple of years till the line between being a Tech Lead and Manager blurred!

When I started my own company, I had the responsibility for building the first product which organically translated into VP of engineering. I enjoyed the job ever since and I often switch between CTO and VP Engineering role on as needed basis!!

What are the biggest challenges you face?

One of the biggest challenges I face is hiring. It requires sustained effort to source and bring on the right talent to help drive our research and product development forward. Once we’ve onboarded any new candidate, it is important to get on the same page about team and company-wide goals. Right out of the gate, we need to be aligned so that we’re all marching toward the same objectives across teams.

At Kindred, we’re driven by results. We want our team members to aim for personal and professional progress and feel that individual excellence deserves to be recognized. We celebrate both individual and team wins, but also know how important it is for our people to maintain healthy work-life balance. As a manager, it’s important to me that my team is happy and fulfilled and I believe this balance, in turn, cultivates a strong collective work ethic.

Read the full interview here

Kindred introduces SenseAct, the first reinforcement learning open-source toolkit for physical robots

 
dxl-reacher-trpo.gif
 

BUSINESS WIRE (SEP 2018) - Kindred, an AI and robotics company that builds intelligence for robots, has announced today the launch of SenseAct, the first open-source toolkit to set-up reinforcement learning tasks on physical robots. Kindred’s SenseAct was created to provide robotics developers and researchers with a consistent, learnable interface that efficiently controls for time delays, a factor that simulation environments aren’t hindered by.

“SenseAct is an important new step in machine learning research on robots, enabling consistent and reproducible results on physical robots for the first time. It will establish a standard that may greatly accelerate machine learning research on physical robots, pushing reinforcement learning to a new level of performance just as large standard data sets have for supervised learning,” said Richard S. Sutton, professor of Computing Science and AITF Chair in Reinforcement Learning and Artificial Intelligence at University of Alberta.

SenseAct allows reinforcement learning agents written for OpenAI’s popular “gym” simulator to learn behaviors for robots, by insulating them from the complexity of real-time control of robotic components. SenseAct’s guiding principles of minimizing delays and maximizing timing consistency via proactive computation lead to responsive learned behavior and reliable learning via state-of-the-art algorithms.

“This initial release focuses on tasks that explore timing, control frequency, action representation, partial observation, and sparse reward. We intend to update SenseAct as we define new tasks to reflect challenges in developing efficient, general, intuitive behaviors for our current and future products,” says James Bergstra, Head of AI Research at Kindred.

SenseAct allows general reinforcement learning algorithms to learn diverse tasks on diverse robots, and is made freely available for anyone to explore and extend. For more information, visit www.kindred.ai/senseact

The Engineering People Show, Episode 25: Dr. Anurag Maunder

I know some engineers will be disappointed by that, but it’s easier for me to find a smart person than for me to find a hardworking person. If there is one message I can give to every young engineer is that don’t take working hard for granted.

Today on Engineering People

Dr. Anurag Maunder is an SVP of Engineering at Kindred AI, an artificial intelligence company that solves human problems with autonomous robots. He started his career with prestigious AT&T Bell Labs. He was the Founder and CTO of Kazeon Systems (later acquired by EMC) and the Founder CEO of dLoop (acquired by Box). In the last few years, he established the Advanced Development lab for EMC to focus on Machine learning for automating storage configurations, and also created Johnson Controls Innovation Garage, which introduced several new AI products in industrial IOT. His current areas of interest are AI, Robotics, Smart Buildings and IoT Security using Blockchains. He believes that AI and Robotics together will fundamentally transform the quality of life and disrupt every single industry.
(Listen to the original episode released on September 10, 2018)