Cynlr 1
Cynlr 1
C O N FI D E N TI A L
2023
Recruitment
Exercise
C O N FI D E N TI A L
Contents
What is CynLr about? ............................................................................................................................................................. 2
Skillset Definition:..................................................................................................................................................................... 3
Evaluation Process: .................................................................................................................................................................. 4
Evaluation Criteria:................................................................................................................................................................... 6
How will your Career Progress? .......................................................................................................................................... 7
C O N FI D E N TI A L
Object Intelligence
For Robots
Expanding AI from Words to Objects
Today, even a simple task of putting a screw into a screw hole without slipping a thread remains
non-automatable across the globe. Imagine what it would take for a robot to assemble a
Smartphone or a car (Musk’s words) by putting together 1000s of parts of varied shapes and
weights presented in random orientations. Enabling robots to perform Picking Orienting & Placing
of objects straight from its container has been a long-standing unsolved problem – touted as The
Holy Grail of Robotics and has remained unsolved over 40 years.
We at CynLr, build the missing pieces of Visual Intelligence for Robotic Arms to make them
Object aware to manipulate with superior agility - adapting to the varying shapes, orientations
and weights of the object – (a demo video link). This we envision could simplify and standardize
large, customized factory lines into LEGO blocks of micro -factories.
C O N FI D E N TI A L
Skillset Definition:
The purpose of this evaluation is to identify resources who can fill the need for the below illustrated
combination of/particular Engineering skills that our product development demands. Treat this
evaluation process as a practice and self-exploration, rather than a typical job interview.
As this requirement is for a Unique Deep Engineering skillset which is not a commonly practiced
occupational role in the market, these skillsets can be augmented either as a single role or multiple
C O N FI D E N TI A L
roles, which is entirely based on the individual’s proficiency with the list of skillsets. This evaluation
process is entirely focused towards objectively identifying those proficiencies.
The individual must be proficient with at least one from each of the above categories -> minimum
of 1 in Principles Understanding ‘&’ 1 in Thought Process / Implementational Knowledge.
This Evaluation is Tool Agnostic and someone’s proficiency or practice with any particular design or
Programming tools like MatLAB, LabVIEW, Python, C++ or any specific Libraries like OROCOS, PID
toolkits, etc is not the priority. Ability to design and build the “fundamental Engineering Principles”
behind those Libraries and Tools is our seeking, which anyways will demand knowledge of some
corresponding tool but not limited only to that.
Ability to Think clearly to know your assumptions (critical thinking), fascination for the problem and
learning quickly matters more than the authority of presumed ‘academic’ knowledge.
Evaluation Process:
The Evaluation journey intends and strives to give an experience of the work that the candidate
would be performing at CynLr, to help the candidate assess the job better - before deciding.
Questions that will be asked, including the Online MCQs, is something that we face in our day-to-
day Product and technology building and not some derivation of popular Interview puzzles.
All questions test the thoroughness of Thinking with the Fundamentals rather than the ‘mastery’ in
implementing a borrowed method / knowledge. (Not the usual competitive/academic/coding
challenge topics)
C O N FI D E N TI A L
C O N FI D E N TI A L
Evaluation Criteria:
1. Critical Thinking: Access to Information is not any more a challenge. Knowing which amongst
the ocean of information is relevant to the problem at hand is the biggest challenge that we
face. Its only made harder, as we are constantly immersed in opinions and biases of brands
masquerading as facts, often assuming the “trust” of authority that the brand holds on a subject.
Yet we can’t ignore them either. That can be tag names like Google, Research of MIT, that Journal
and this Scientist, etc. and the worst of all - ourselves. We tend to put too much of “trust” on our
own first thoughts and perceptions. Problems don’t come announcing and we realize a problem
only through our first perception and we build from there, which is often a bundle of assumptions.
Yet we are more likely to be sure of what we “feel” and mistake them as our thoughts. So much
so that, we use logic to justify our perceptions rather than to validate our perceptions.
Critical Thinking is the fundamental thought culture that we seek in every role that we play at
CynLr. To know our assumptions is Critical Thinking.
3. Design Thinking: (Since this is hard to objectively evaluate it will be a relative evaluation both
with the other submissions and with the self across consecutive submissions) Purpose of every
design is to minimize human effort towards solving a problem. Which beckons us to see how
all this demand for human effort manifests itself. It begins with human inabilities and limitations
- Fatigue, sensory limitations, sensory Deception leading to Confusion (eg: people banging on to
a transparent glass door), human misconceptions, misinterpretation, misperceptions, misuse,
misapplication – all put in one bracket as Human Error. Adding to it is the innate human desire
to minimize effort, which easily distracts them to be bored – lack of foresight, or worse- urge to
tinker. All these are a common factor for the person who devices the solution and for the person
who uses the solution. A design must factor all the above and must try to minimize and cover as
much as possible for various possible scenarios.
C O N FI D E N TI A L
Training/ Induction as
Learning Probation
Organizational Contributor
Of course, then the thought process that we require are not readily available or academically groomed
and therefore requires a significant unbiased learning about the problem before learning the skill to
solve it.
We often end-up spending more effort in unlearning the misconceptions and assumptions that typical
job training infuses– the usual ML, Computer Vision, etc, mostly calibrated towards building talent to
fit the widely needed jobs. Such jobs augment solved problems that need scale and not necessarily
unique, concentrating more on the ‘tools’ than the fundamentals – pick a library and figure out how to
integrate code blocks or some mechanisms or circuits and make an application .
If this approach would have worked for our problem, then the problem would already be
solved. Hence it demands a foundationally unique approach towards to solve Object
Manipulation. To quote, CynLr builds its own “TensorFlows” and “OpenCVs”.
This is endemic not just to the SW and ML algorithms, but to almost every HW layer & Infrastructure
used – Image sensors & optics to Haptics & Arms to Computational Platforms to even Lab Facilit y &
Supply Chain, had to be uniquely designed from scratch while pushing their limits.
In this context, the candidate can expect to help CynLr build organizational capacity in all kinds of
Design from Research to Product Development to Brand Building to Operations to Applications &
Sales or diversify or specialize in any of these domains.
C O N FI D E N TI A L
ssss