A lot of my projects are interlinked. I have created an interactive visualisation to have a overall picture of my work. I have made a list view of all my projects. Please click here to view them.
Click the circle to know my involvement with a particular topic and the click the branch link for the corresponding project linking the two topics. The branch will change its color to red when you hover above it.
I majored in Physics at Indian Institute of Technology (IIT), Roorkee and graduated in 2014. I feel happy and proud to have studied there. During my high school time, I wanted to study at IIT and I'm glad I realised my dream after cracking IIT-JEE. I chose Physics to have a good understanding of how things work in the world. The physics approach of thinking always helped me to see through things better and gave me an edge when I worked with my peers from Engineering and other backgrounds.
During my years at IIT, I learned a lot, both academically and otherwise. IIT Roorkee boasts one of the best sports infrastructures I enjoyed playing Tennis and hiking trips. I was interested in optics and did my best to spread knowledge of this subject. I started a student chapter of the Optical Society of America at IIT Roorkee and organised gave lectures and demonstrations to the nearby schools.
|2017||Self Driving Cars||Udacity||Completed|
|2014||IMSc Physics||IIT Roorkee||CGPA:6.88/10|
|2008||Intermediate||MVM School, CBSE Board||
|September '18- Present||Machine Learning Lead||Ayu devices|
|August '16- Present||Owner||Vikraman Karunanidhi Consulting Services|
|2015 JULY-FEB '16||Okinawa Institute of Science and Technology(OIST), Japan||Light Matter Interaction||Research Technician|
|2015 MAR-JULY||Indian Institute of Technology(IITD), Delhi||Photonics research Labarotory, Physics department||Research Internship|
|Indian Institute of Technology(IITM), Madras||Plasmonics research, Electrical Science Department||Research Internship|
|Indian Institute of Science(IISc),Bangalore||Soft Nano Materials lab, Physics department||Research Internship|
I won an award for the best Innovative solution at the Vivli-Microsoft Datahon. Participants were challenged to present a method to safeguard data of patients suffering from rare-health diseases.
I represented my firm at the Maharastra AI Innovation challenge(Mumbai, March, 2019) and won top spot in the healthcare. The event was organised by Government of Maharastra, Niti Aayog. Winning at this event, gave us an opportunity to deploy our solution in Hospitals and public healthcare centers in Maharastra.
I secured the top 1%(out of 6730 students) in the National Graduate Physics Exam(NGPE) in the year 2011. NGPE is a national level exam for Physics students in India.
I was an active member at Electronic section, IIT Roorkee and made several prototypes. 'Srishti' is the Annual Hobby Fest of IIT Roorkee,(annual exhibition of IIT Roorkee)
Won 1st prize Gaming electronics for "Car controlled by hand motion" during Srishti-2011
Won 2nd Prize for "Automatically Aligning Solar Panel" during Srishti-2010
I was a recipient of the prestigious Innovation in Science Pursuit for Inspired Research(INSPIRE) Scholarship(2009-14) awarded by Ministry of Science Technology, India for academic excellence to support my studies and research internships.
I was among the top 10% scorers from my school in National Mathematics Talent Competitions (NMTC) in 2005, organised by the Association of Mathematics of India(AMTI). I got a chance to attend a Mathematical workshop. To aim of this exam was to discover and encourage students who have the capacity for original and creative thinking, readiness to attack unfamiliar and non-routine problems exhibiting a general mathematical ability appropriate to their level. ( link to website )
I received 3rd prize in a painting competition held by Vision India (in 2003) and also won my first cash prize of Rs.250/-
"A deep neural network to identify foreshocks in real time" by K.Vikraman link to paper
Annual Seismological Society of America(SSA) Conference(2019)
A deep neural network to identify foreshocks in real time (Oral presentation)
Event and noise discrimination using AI (Oral presentation)
Applied Optics and I :
During the second year, I took up my first project in Applied Optics to improve light extraction efficiency of OLED and started reading more about the Plasmonics field that connects Physics and electronics at a nano level. My interest in optics augmented. To have more lively discussions, and expand my activities, I started a student chapter of the Optical Society of America in my institute. I held active talks about Applied optics and gave a few demonstrations in a nearby school. In the coding front, I gained experience about powerful numerical computations like MEEP FDTD and Lumerical FDTD and made some tutorials and videos on them. For my final year thesis, I worked on moth-eye structures to reduce reflection loss in Solar PV. After graduation, I did a three-month internship at IIT Delhi, working on biosensors. I designed a novel biosensor based on WGM, that has narrow bandwidth and good sensitivity. From July-15 to Feb-16, I worked as a research technician at Light-Matter Interaction Unit at OIST, Japan. Majority of my work revolved in fabricating nanoscaled devices in the clean-room. I successfully fabricated a nano bridge using FIB machine. In July-2016, I was lucky to participate in 'Engineering the eye'' workshop where I worked on the device 'Julia'. My contribution towards the project was that I redesigned the device simplifying the Optics.
Deep learning and I :
My first encounter with Machine learning was in 2014, when I did a course on Machine learning by Andrew Ng. Almost after a year, I came in a close encounter to major earthquakes. The news that thousands of people dying because of earthquakes was disturbing me. As a Science graduate, I questioned why earthquakes are unpredictable? Coincidentally, there was a new reading club formed at my workplace(OIST, Japan) to discuss the about Deep learning book. I actively took part in the discussions and started learning more about this subject. After completing my project in Japan, I decided to devote more time to self-research to find a reliable precursor. I started learning more about seismology in-depth, analyzing past methods of prediction. After a long search, I created an algorithm that takes in live data and identifies whether it is a foreshock, mainshock or an aftershock. During the same time, I got selected for an elite nanodegree program about self-driving cars by Udacity. I worked in the following fields: Computer Vision, Deep Learning and Sensors.
Healthcare and I :
My late grandfather was a famous doctor in a remote village. Growing up, I saw how local people used to worship him and his service to the community. He will always remain my role model for my endeavours in the medical field. 'Medical Device Innovation Camp(MeDic)' at IIT Bombay and 'Engineering the eye' workshop conducted by LVPEI-MIT Media Lab exposed me to practical aspects of how technology can help solve medical woes. I currently work at a healthcare startup, that allowed me to interact with the best cardiologists in India. I've participated in a medical(Kaggle) competitions(Pneumothorax Segmentation,Pneumonia and Nuceli detection) and firmly believe that AI can mitigate the doctor-population ration issues.
Image processing and I :
My first experience with Image processing was in Robotics competition held at IIT Bombay,2011 which involved a dynamic areana. After that, I made some algorithms to track hand gestures and control a small car with hand gestures. In May 2011, I analysed images took from AFM at IISc using image processing algorithms. Now I use deep learning algorithms to study and classify images. During my self-driving car course, I worked extensively on Computer Vision algorithms for detecting lanes, cars, traffic signs and behavioural cloning.
I have created schematic diagram to visualise my projects. To toggle there, click here.
I created an algorithm based on deep learning that can discern foreshock, mainshock or aftershock in real-time. Until now, it is impossible to identify the type of earthquake in real-time. Identifying foreshocks as and when
they occur provides valuable insight to predict imminent earthquakes. After training, I successfully classified them into their respective types with high accuracy (>99%). The project is one of the first applications of deep
learning in the field of seismology, and I am thrilled and excited about it.
link to paper
One of the primary cause of blindness is due to refractive errors. All the available devices to detect them are pretty expensive and bulk. A team at Srujana Innovation centre, which has a tie-up with MIT Media Lab came up with a solution called "JULIA"-Just use logic for Imaging Applications. The current version had shortcomings in measuring power, some ergonomics and necessity to focus at an image 3m away. I got a chance to work on the latest iteration during the "Engineering the eye Workshop" conducted at LV Prasad eye institute from 3-9th July, 2016. I worked on the optical design and together as a team, we redesigned Julia from scratch to a stand-alone unibody design. Using big data approach, we incorporated human errors making Julia readings reliable. The best part is that the total cost per device can be scaled to Rs.20/-( few cents)
The goal of this project is to fabricate SiN beams and couple the mechanical vibrations of the beam to a microcavity and study their effects. The main challenge in fabrication is that SiN membranes are fragile, making it difficult to fabricate. After several failed attempts, I fabricated a good quality SiN beam using Focussed Ion Beam(FIB) method. I learnt different fabrication techniques like mask-less lithography, Inductively Coupled Plasma(ICP). I made simulations using Comsol MultiPhysics to design the structure in order to obtain the desired frequency.
I developed a new design for sensor-based on Guided Mode Resonance(GMR) effect with high sensitivity of 430nm/RIU and narrow FWHM of 0.05nmthan conventional GMR sensor. With high sensitivity, my proposed design finds
application in detecting biomolecules through refractive index sensing.
I participated in MIT Media Lab's Design and innovation workshop(2014) held in Mumbai, India. I, along with a team of four members, developed a prototype for recognizing hand gestures aesthetically. We used a smart fabric whose resistance is pressure sensitive. We knitted the fabric into a glove and measured the changes of pressure using an Arduino. We could discern between different hand gestures. Potential applications include emergency call, Human-Computer Interactions and so on. From this workshop, I learned a lot about brainstorming and making a good prototype.
One of the significant losses hindering the efficiency of the solar panel is the reflection loses. There is a considerable difference in refractive indices between air and glass interface which accounts for a loss up to ~45% of the incident light. I took this problem as my master's thesis topic. It is known that Moth's eyes are made of nanostructures that reduce reflection losses. My goal was to optimise moth-eye structures for Solar PV and increases. I was able to achieve low reflection loss less than 2% for a broad range of wavelength for different angles of incidence.
This was my first research project. This project was a colloboration between Physics department and Moeser Baer company. The goal of this project was to improve extraction efficiency of OLED. The light coming out of a OLED device is only 20% of what is produced. This is due to reflection,transmission and other losses. To combat these losses, we used Silver nano particles. My contribution to the project was to study the effect of size and shape of Silver nano particles and their effect to improve the light extraction efficiency of OLED.
I made a prototype to identify hand gestures and used them to steer a custom toy car. To recognize hand gestures, I attached a colour tag in my forefinger, which my camera was tracking. I wrote a code in MATLAB to monitor the colour tag and interfaced and I interfaced it with an Arduino. The MATLAB code provided instructions to the Arduino. In turn, the Arduino controlled the motors and steered the car accordingly. I exhibited my prototype during IIT Roorkee's annual hobby fest and received 1st prize and accolades from the audience.
In general, the position solar panels are fixed. However, more amount of solar energy could be harvested if we can position the solar panel in the direction of maximum intensity of the sun's rays. To solve this issue, I made a prototype to track the path of maximum intensity of sun's rays. I used a phototransistor, Arduino and two servo motors to find the direction(both polar and azimuth angles)accurately. For more details on the working, check out the above video.
Deep learning frameworks: TensorFlow, scikit
I have my Youtube channel where I discuss code and algorithms implemented in TensorFlow
I support python and contribute to the growth of it by writing articles in my blog. I love the graph options in R and currently use R Studio for data analysis.
I studied C++ in high school and got 89% in my computer science in my 12th standard board exam. I completed a course on Computational Physics, where I used MATLAB for my assignments. I used OCTAVE in my Machine learning class by Andrew Ng.
I am proficient in the CAD module in Comsol. I used them to simulate optomechanical structures. I used Lumerical FDTD to model and study effect of Motheye structures in Solar PV and also in designing a new biosensor.
MIT Electromagnetic Equation Propagation(MEEP) is based on scheme language. I used it during my studies in Photonics structures. Although MEEP is a powerful tool, it currently has limited resources for beginners. I support and promote the MEEP community by writing blog posts. One of my popular blog posts about MEEP is "Getting started with MEEP" . Some of my simulations in optics.
Gimp and Inkscape
During my research work at the Light-Matter Interaction, OIST, I used Solid Works to design custom mounts for different optical components. In my previous research experience at IIT Delhi, I used OpenScad to create optical mounts, and 3D printed it.
Web development Skills :
I am profienct in the following web development languages. I used them to buid my personal website.
HTML & CSS
D3js (Data Driven Documents)
Arduino and Raspberry Pi
I was actively a part of the electronics section, IIT Roorkee, where I learned about breadboard prototyping, soldering and programming in Arduino and Raspberry Pi. I support and promote the use of Raspberry Pi through blog articles and videos.
I have received formal training and I have worked with the following equipments/process
Scanning Eletron Microscope(SEM)
Focussed Ion Beam