About the Role:

Robotics is an industry that has been gestating for 70 years yet it has never arrived.

Every time, its perceived moment arrived, it was always an Use-Case for some other technology, including now - an Use-Case for #LLM/VLA. Intelligence is a Slave of Data-> Slave of Signal -> Slave of Sensor. Not Action. Robotics is Formfactor and action obsessed, but autonomy begins from Sensing.

Robotics is often mistaken for a problem that's just lacks Savvy Engineering to put together current technology in a meaningful way. In reality - Robotics lacks & needs foundational and fundamental technology & tools needed to Build it. This is a Foundational research problem.

Robotics has a problem of Building a Car with Cycle Wheels. None of the technology or the components that we use for robotics were developed for Robotic Manipulation.

Beyond the dancing mannequins, Robotics is a Foundational #Research problem, and it begins from Sensing by Action to Force Controlled Motion than just Position Control to Multi-Arm Action Planning to Foundations of #ML that Mirrors #Neuroscience of Cognition & Motion.

! .

If you are a researcher, a PHD student, Post doc or a Research scientist/ Faculty and if you connect with the Constraints of Current Approach to Intelligence, Sensors, Compute and Robotics and is eager to dive deep into the Rabit hole - Apply.

https://lnkd.in/gDCNtJnw

Research Topics & Domains:

Vision and Algorithms

Dynamic Vision - Vergence, Motion Imaging & Dynamic depth – Hierarchical Depth Construction, Imaging, Optics & sensors – Building intelligence into Sensing
Exploring primitive Neural layers/ Pathway of Human vision
Object Composition – Texture, contours, grasp mapping, force field – process of learning an object - Primitive Visual Patterns and stitching of contours into an object
Process of Colour Construction
Understanding Reflections, Attention & Priority of Feature Selection
Visual Models - breaking down scenes into Action models
Involuntary & Voluntary Imaging

Manipulation & Sensing

Skin Transducers for Touch – RF, Capacitive, etc - Modular Tactile Design for the shape and format of Fingers & Arms
Convergent Camera Design & Image sensor design , Autofocus Liquid Lens Optics Design
Neuromorphic Image Sensor pipeline - FPGA
Sensing Primitives- Creating Sensor Physics in Simulation Environment to map Sensor's signal response for different object interactions
Alternate Modes of Actuation for Robots & Motion Control – Tendon, Tendon + Joint Actuation
Motor & motion Design for Wrist and Ball joints

Machine Learning

Neuromorphic NN designs - Neural morphology - Breaking down neural patterns and their Functions
MPC & RL Design
On-the- Fly Learning process & Live weight modelling
Signals to senses, closed loop architectures in Neurons
Adaptive Learning Architectures – Dynamic Weights, Static & dynamic neural patterns
Neural memory & temporal neurons
Dimensionality of physical actions encoding into neural pathways
Auto Object learning Setup for Robots

Robot Control

Multi arm Kinematics Multi Arm- task Scheduling and Task sequencing
Grasp Models And Force Modelled Grasping Algorithm Library for
Gripper Design & Construction
Multi-Finger Trajectory & Grasp Planning
Action primitives – Action models
Data (Signal) Model of Objects
Dynamic Collision planning
Energy Optimized path Planning and motion breakdown between Arms Force- Primacy Control
Breaking down tasks into Micro Actions and Building Micro RL models for Grasp, Task

Requirements:

What do you need to thrive in this Job

Business empathy - Most important mistake that Engineers and Researchers make is Not including the constraints of business as a factor of the problem we solve. From Money, volatility of customer interest, Runway, resource availability, Supply chain bottlenecks, Manufacturability, Scalability, Serviceability, Team Alignment, tool, Skill & infra bottlenecks – all of these dictate the relevance of what one builds more than the principles and tech. A research that doesn't track its relevance to these metrics most often ends up unviable. Miss one, the whole approach goes unusable or unviable.
Outcome oriented Research – Key factor that differentiates Academic research from Industrial Research. Have developed structured methods to accelerate open ended problems
Customer/User focus – This is the north star and gives purpose. The greatest joy is in seeing someone benefit from your work. Even greater joy is in they getting addicted to your work. Not Paper focused.
Discipline & Consistency - Research is an unobjective business and its easy to hide behind its subjectivity and justify delay/ mood swings. When system lacks objectivity, Its self-discipline that gets you to the goal. Timelines Matter – Simply because money runs out. Easiest hack – Learn to carve out usable outcomes our of every step of progress in Research, even if the entire research is yet to materialize.
As part of the Job you will be responsible for joint research efforts with Research institutions including IISC & EPFL. You will be responsible to coordinate the research activities between CynLr & the Partner Research Institutions. And manage outcomes out of the engagements.

Who should Apply

If you are a Post-Docs, Phds, Research faculties, Research scientist looking to further your research into materialized real-world outcomes
You are in your UG/PG, but beyond the romance of research & titles, if you have understood the grind of research and is keen on building technology that needs solving a foundational problem for customers/industry that requires research, please apply.
Driven Not by Papers, Not Patents, Not titles – just utility, Please apply. Understand the difference between experimental research and paper research.
If you have realized the limitations of current approach to Intelligence and Robotics technologies and recognize the need for deeper research.
Structured research mindset – Not Jugaad.

Technical Skill :

Very strong fundamentals understanding in respective subjects.
For Algo & Compute – HPC/GPU role, you need to have exceptional understanding of Computer Architecture, Compliers & OS.
For Computational Neuroscience & Vision - Proficient with Computer Vision, Physics of Light & Optics, Fundamentals of Neurons/Neuroscience.
For Optics, Sensors & HW - Proficient with Electronic Design, Materials, Optics, Signals & Systems. Savy with instrumentation.
For Robotics & Motion Control – Strong with Controls Theory, MPC & Adaptive control, working experience with Motion Control Design, Proficiency in Robot Control & Kinematics, Machine learning Fundamentals.
For ML/RL research – Proficient with ML Architecture design, Temporal NN design, Designing Reinforcement Learning , Foundations of NN and neuron design, Learning Design.
Fundamentals of Systems and Control theory
Strong penchant for Instrumentation (designing measurement tools/methods & sensors), analytics and designing experiments
Proficiency with programming– You don't have to be tool/Language experts.

What This Role is NOT

Pure academic paper driven research.

Team Structure

The Technical Program Manager operates within the Founder's Office, embedded directly into the engineering organization.

While the TPM works closely with the founder(s), the role interfaces continuously with:

• Engineering leads

• HR and Talent, for performance programs and people-related interventions

• Finance and Operations, for resource planning and project budgeting

• External vendors and tooling partners, where applicable

The TPM owns process design and delivery hygiene across all engineering streams — acting as the connective layer that keeps multi-disciplinary teams aligned without creating dependency on a single coordination point.

About CynLr

As a foundational technology building company in Robotics & AI, CynLr builds visual robots - Cynoid & CyRo that can intuitively learn to pick & handle even unknown objects without requiring any prior training, just like a human baby fiddles with objects. CynLr calls this stack as Object Intelligence (OI).

From fitting a screw to removing object out of its the plastic wrapper to automating the assembly of a car/gadget - every such object handling task that involves adapting on the fly is not prior trainable & thereby remains non-automatable across the industries. CynLr builds a stack of Micro RL blocks called Sensing & Action Primitives, using which a High Level Task can be composed.

With OI's ability to learn on the fly, CynLr's focus is to universally automate factories and eliminate the need for complicated custom machines to manufacture products. Thereby simplifying manufacturing into Universal Factories, which can be programmatically repurposed to produce wide varieties of Products.

CynLr s Vision and ML stacks are deeply inspired by neuroscience, modelling how biological vision understands shape, geometry, and interaction rather than appearance alone. To support this, CynLr builds its hardware, sensors, compute pipelines, and learning stacks from scratch, tightly coupling perception, decision-making, and action.

We are a deep-tech robotics company at the intersection of computer vision, real-time control systems, and precision hardware, headquartered in Bengaluru with a Swiss entity. Our engineering spans optics, embedded systems, firmware, real-time software, mechanical design, and applied research.

Member of Technical Staff - Research