About Me
I am currently working in the Emerging Technologies team in PwC US Advisory. The team focuses primarily on providing innovative, cutting edge solutions to client problems leveraging AI, ML and IoT. In addition to client projects, we also build PoCs (Proof of Concepts) for different technologies which we think will bring value to clients and present them as potential replacements for outdated ways of working.
Before this, I did my Masters and Bachelors in Chemical Engineering from IIT Kanpur. I was interested in Simulation studies and did multiple projects in this domain including molecular dynamics simulations, simulating flows of particles and simulating burner profiles. All of this was built leveraging multiple open source packages and products along with self written optimised code.
Apart from this, I like board games, online strategy games and geeky tech stuff. I also like teaching and have been pursuing several opportunities to teach underprivileged students. I love keeping up with the latest trends in technology both in the hardware and the software domain.
Download my resumé.
- Vision
- Robotics
- Artificial Intelligence
- Hyperconnected networks
- Computational Science
-
M.Tech. in Chemical Engineering, 2019
Indian Institute of Technology Kanpur
-
B.Tech. in Chemical Engineering, 2018
Indian Institute of Technology Kanpur
Experience
I am currently a Senior Associate in the Emerging Technologies team. I was given an EARLY PROMOTION for consistently being the best performer in the team for 2 consecutive years.
My current responsiblities include :
- Manage operations of the emerging tech team and spearhead initiatives across organisation
- Lead projects from ideation to completion in stipulated timeframe maintaining delivery quality
- Come up with ideas for different PoCs that can be pursued to improve the overall offerings of the team
- Solve client problems with innovative, cost-effective and easy to deploy solutions generating value
- Stay in touch with the latest & best and work on the convergence of AI, ML and IoT in client offerings
Some of the projects I worked on:
- Client - Smart AI - Hospital Code Correction - built a tool to analyse medical claims for hospital diagnostic (Dx) and procedure (Px) codes correction in final billing at discharge
- Client - Contract Lifecycle Management - extracted information from legacy contracts, built taxonomoy to categorise them and linked contracts hierarchically
- Research - Stylized NFT Generation - built a serverless deployment on azure functions to stlyize images based on templates for NFT generation and marketing
- Research - Deepfake Audio Generation - explored zero-shot deepfake audio generation models to generate audio from text using embeddings of short reference audio clips
- Research - RL Alternative using ES and DL - explored an ensemble of deep learning and evolutionary strategies as an RL alternative in the context of an investment problem
- Research - Model parallelism for inference at edge - explored distributing model components across different devices to enable running large models on edge with multiple smaller devices
- Research - ReadTheDocs Documentation Templatisation - explored sphinx and created a solution to automatically generate github-hosted documentation under 10 mins in RTD format
- Firm Dev - Campus Recruitment Lead - led recruitment process for my alma mater, IIT Kanpur
- Firm Dev - Team Connectivity Lead - led the team organising fun events and team building activities
I worked as an Experienced Associate for about 1.5 years in the Emerging Technologies team.
My responsibilities included :
- Lead workstreams in projects and get them completed within the timeframe
- Come up with ideas for different PoCs that can be pursued to improve the overall offerings of the team
- Solve client problems with innovative, cost-effective and easy to deploy solutions generating value
- Stay in touch with the latest & best and work on the convergence of AI, ML and IoT in client offerings
Some of the projects I worked on:
- Client - Spatial AI - analysed and predicted the behaviour of visitors in NYC’s Bryant Park, building a digital twin spatial model for 4D simulations
- Client - Occupancy Detection - built serverless solution for anonymised occupancy detection inside office space using LIDARs and WiFi sensors with social bubble breach detection and occupancy tracking
- Client - Factory Intelligence - built a smart factory solution with unplanned machine downtime detection and industry-standard OEE calculations for key KPIs using cost-effective sensors
- Research - Concept Drift - explored concept drift and built an sklearn-like toolkit to detect drift in image, text or audio data using multiple drift detection methods for extending life of production models
- Research - Automated PPT content editor - built a tool that transforms any PPT to a formatted PwC-compliant file including several Editorial, Branding and Alignment changes using lxml open office format
- Research - Distributed Edge Compute - built and deployed distributed edge compute on edge with fractional device usage for higher device efficiency using load balancing and micro kubernetes engine
- Research - Demand Prediction with competition analysis - predicted demand using sales figures, mobility data and several demographic and footfall variables
- Research - Industrial IoT Platform Comparison - Compared different Industrial IoT platforms from a technical and business perspective and evaluate on different metrics for adoption
- Research - Fingervein Detection - designed end-to-end biometric system with user registration and verification as a replacement for fingerprint technology
- Research - IoT Platform Comparison - Compared the IoT offerings provided by major cloud providers from a business perspective and evaluate on different metrics for adoption
- Hackathon - ImageCLEF DrawnUI 2021 - $2^{nd}$ Rank - extracted HTML UI elements from wireframe drawings of websites and website screenshots using YOLOv5 and confidence cutoff variation
- Hackathon - Twitter Bias Hackathon - $9^{th}$ Rank - highlighted bias instances in Twitter’s Saliency model used to crop images building a case study with images from Tokyo Olympics 2020
- Hackathon - NEURIPS L2RPN 2020 - Robustness Track - $28^{th}$ rank - RL-based challenge to robustly maintain an electrical grid without disruptions against an adverse agent
I worked as an Associate for about 1 year in the Emerging Technologies team.
My responsibilities included:
- Solve client problems with innovative, cost-effective and easy to deploy solutions generating value
- Stay in touch with the latest & best and work on the convergence of AI, ML and IoT in client offerings
Some of the projects I worked on:
- Client - Asset tracking - tracked cargo shipment using cost-effective sensors along its cross-country journey from factory to distribution center ascertaining possible locations of damage
- Client - Damaged car parts segmentation for auto claims - performed damage segmentation for automating auto claims with explainability and automated claim / damage report generation
- Research - Active learning - explored Active Learning with different querying strategies on CIFAR10 dataset and managed to achieve high accuracies with very limited training data
- Research - Implementation of live models on edge - implemented and optimised multiple live models (eg. action recognition model with over 200 classes) on edge devices (eg. Jetson TX2, Raspberry Pi)
- Hackathon - ImageCLEF DrawnUI 2020 - $3^{rd}$ Rank - extracted HTML UI elements from wireframe drawings of websites ideating novel multi-pass inference technique to boost recall
- Firm Dev - Training Tracker Repository - ideated and built out a repository management system for all internal trainings to boost self-learning and incorporate feedback into future trainings
I worked as an intern for 2 months at TRDDC Pune. I completed the work well within the time frame and had time to complete an extension of the original problem statement as well. I was awarded an Excellent grade for the internship for immaculate coding skills and dedication to work.
As an intern, the projects I worked on:
- Macroscopic model of upper layer of skin - explored and generated a macroscopic model of skin’s upper layer ‘stratum corneum’ using OpenFoam, a tool being used for the first time in the facility
- Effect of electroporation on transdermal drug delivery - introduced electroporation in macroscopic skin model and simulated electroporation which were found to be in agreement with experimental data
Patents & Publications
Served as a Program Committee Member / Reviewer for the following conferences:
- 3$^{rd}$ International Conference on Artificial Intelligence and Machine Learning (CAIML 2022) - Toronto, Canada
- International Conference on Artificial Intelligence Advances (AIAD 2022) - Virtual
- International Conference on Machine Learning, NLP and Data Mining (MLDA 2022) - Virtual
Projects
Contract Lifecycle Management (CLM)
Extracted information from legacy contracts, built taxonomoy to categorise them and linked contracts hierarchically
Zero-Shot Deepfake Audio Generation
Explored and created a web-hosted demo for a zero-shot deepfake audio generation model to generate audio from text using embeddings of short reference audio clips
Intelligent Systematic Investment Agent
Explored an ensemble of deep learning and evolutionary strategies as an RL alternative in the context of an investment problem
Model Parallelism for Inference at edge
Explored distributing model components across different devices to enable running large models on edge with multiple smaller devices
Read-The-Docs Documentation Templatisation
Explored sphinx and created a solution to automatically generate github-hosted documentation under 10 mins in RTD format
Concept Drift Detection with Chatbots
Explored concept drift and built an sklearn-like toolkit to detect drift in image, text or audio data using multiple drift detection methods for monitoring and extending life of production models
Spatial AI
Analysed and predicted the behaviour of visitors in NYC’s Bryant Park, building a digital twin spatial model for 4D simulations
HTML UI element extraction
Extracted HTML UI elements from wireframe drawings of websites and website screenshots using advanced image pre-processing and confidence cutoff variation. Achieved $2^{nd}$ rank in the hackathon.
Automated PPT content editor
Built a solution that transforms a non-formatted PPT file to a formatted PwC-compliant file including several Editorial, Branding and Alignment modules using lxml open office format
Twitter Bias Hackathon
Participated in the Twitter Bias Hackathon and highlighted bias instances in Twitter’s Saliency model used to crop images building a case study with Tokyo Olympics images as example. Achieved $9^{th}$ rank in the hackathon.
Distributed Edge Compute
Built and deployed distributed edge compute on edge with fractional device usage for higher device efficiency using load balancing and micro kubernetes engine
Occupancy Detection
Built a solution for anonymised occupancy detection inside office space using LIDARs with features including social bubble breach detection and zone-based real-time occupancy counts and tracking
Demand Prediction with Competition Analysis
Predicted demand using sales figures, mobility data and several demographic and footfall variables
Factory Intelligence
A smart factory solution with unplanned machine downtime detection and industry-standard OEE calculations using cost-effective sensors and a dashboard with key KPIs
L2RPN Hackathon 2020 - Robustness Track
RL-based challenge to robustly maintain an electrical grid without disruptions against an adverse agent. Achieved $28^{th}$ rank in the hackathon
Fingervein Detection
Designed end-to-end biometric system with user registration and verification as a replacement for fingerprint technology
HTML UI element extraction
Extracted HTML UI elements from wireframe drawings of websites ideating novel multi-pass inference technique to boost recall. Achieved $3^{rd}$ rank in the hackathon.
Cross-country Asset Tracking
Tracked cargo shipment using cost-effective sensors along its cross-country journey from factory to distribution center ascertaining possible locations of damage for effective preventive steps
Trainings Repository
Ideated and built out a repository management system for all internal trainings to boost self-learning and incorporate feedback into future trainings
Active Learning
Explored Active Learning with different querying strategies on CIFAR10 dataset and managed to achieve high accuracies with very limited training data
Implementation of live models on edge
Implemented and optimised multiple live models (eg. action recognition on something something dataset with over 200 classes) on edge devices (eg. Jetson TX2, Raspberry Pi)
Damaged Car Parts Segmentation for auto claims
Performed damaged car parts segmentation for automating auto claims with additional features like explainability and automated claim / damage report generation
Complete Plant Design of Acrylic Acid
Designed, simulated and optimised a complete plant design for the production of acryclic acid with raw materials on ASPEN HYSYS and MATLAB. Successfully tuned the plant for optimal performance with minimum cost and robustness to raw material stream fluctuations.
Macroscopic model for Transdermal Drug Delivery
Developed a macroscopic model of skin’s upper layer (Stratum Corneum) to study transdermal drug delivery and effectiveness of electroporation on drug transfer.
Numerical Study of SiO2 Burner profiles
Performed a numerical study of SiO2 burner profiles simulating a burner flame and calculating the flame temperature profile and species distribution in the flame.
Numerical Investigation of Nanoparticle Precipitation in Y-Mixers
Performed a simulation study of nanoparticle precipitation in 2D and 3D Y-mixers solving the population balance equation. The study included fluid flow and tracer analysis using both C++ and palabos library.
PID Characteristics
Studied the control characteristics of a process with temperature scaling. Tuned the parameters using PID tuning on MATLAB.
Composite Allegro 2m glider
Mentored and led a team of 5 people to fabricate Prof. Mark Drela’s Allegro glider (Hand Launch Glider) with a 2m wingspan
Centrifugal Pump Characteristics
Studied the control characteristics of the centrifugal pump system. Derived the transfer function for flow rate and discharge pressure from the response of a step change in RPM and verified the same by further experimentation. Also, calculated the optimum RPM operation point for the Pump setup.
ABU Robocon 2016
Designed 2 robots as a part of the ABU ROBOCON 2016 challenge, one being a powered robot and the other being a non-powered robot that will move using non-contact forces provided by the former and steer using line following. The powered robot was equipped with wall following and pole climbing. Achieved $3^{rd}$ rank nationally out of about 105 teams.
Composite Aegea 2m glider
Fabricated a composite glider aegea (Composite Light Aileron Hand Launch Glider) with a wingspan of 2m based on design by Prof. Mark Drela.
Semi-Autonomous Surveillance and Tranportation Robot
Developed an all-terrain vehicle capable of automatic navigation and surveillance using Embedded systems and Google Maps API.
ABU Robocon 2015
Designed and developed 2 badminton playing robots (doubles on a standard badminton court) for the ABU Asia-Pacific Robot Contest 2015. Achieved $11^{th}$ rank nationally out of about 80 teams.
Certifications
- Build Basic Generative Adversarial Networks (GANs) [Certificate]
- Build Better Generative Adversarial Networks (GANs) [Certificate]
- Apply Generative Adversarial Networks (GANs) [Certificate]
- Natural Language Processing with Classification and Vector Spaces [Certificate]
- Natural Language Processing with Probabilistic Models [Certificate]
- Natural Language Processing with Sequence Models [Certificate]
- Natural Language Processing with Attention Models [Certificate]
- Fundamentals of Reinforcement Learning [Certificate]
- Sample-based Learning Methods [Certificate]
- Prediction and Control with Function Approximation [Certificate]
- A Complete Reinforcement Learning System (Capstone) [Certificate]
- Neural Networks and Deep Learning [Certificate]
- Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization [Certificate]
- Structuring Machine Learning Projects [Certificate]
- Convolutional Neural Networks [Certificate]
- Sequence Models [Certificate]
Voluntary Work
I worked as an Academic Mentor in the Counselling Service Team in my sophomore year and taught :
having scored an A* (for exemplary performance) in the course.
I also served as a member of the Overall Academics Team in my junior year and as Overall Coordinator of the Counselling Service in my senior year, responsible for the functioning and management of the student-run body. My responsibilies included managing teams comprising of a total of 200 students at different levels, managing funds, coming up with campus-wise sessions, and organising the Orientation program for freshmen.
I also served as a TA (Teaching Assistant) in 2 courses as a masters student. The courses were :
- CHE 352A : Chemical Process Simulation Lab with Prof. Vishal Aggarwal
- CHE 492A : Unit Operations Laboratory with Prof. Anurag Tripathi
I also taught pre-college maths and science to underprivileged students in the neighbouring village of Kanpur in my freshman and sophomore year. I currently am teaching a group of underprivileged students and helping them prepare for IIT-JEE entrance exam.
Contact
Please feel free to contact me for anything !