About Me
Hi, my name is Shiv and I am a Senior Software engineer at Bloomberg, NY. I completed my Master’s in CS from New York University and have experience working as an SDE at Adobe and SDE intern at Tiktok.
I also enjoy: Designing and Sketching
Experience
Technology Infrastructure
- Developing APIs and workflows for users to launch physical and virtual machines within the company. Led different projects to integrate new features and improve system performance.
- Collaborating with different teams like storage, network, hardware, infrastructure etc; used Airflow and Python.
Developers Portal Tiktok
- Owned the design and development of history management feature in the Tiktok developers portal; in Go.
- Extra project; developed an optimal algorithm for scheduling meeting amongst people around globe; in Python.
Containerizing web-scrapping
- Parallelized Python data crawling using multiple containers, highly appreciated by Professor Raveesh Mayya, NYU-Stern, for doubling scrapping rate with lesser resource utilisation using containers.
- Data extraction and analysis of large online secondary market using Python multiprocessing and RapidMiner.
- Ported the setup to GCP for efficient management and processing.
Also explored management, showed leadership and customer support.
- Owned entire development of Integrations of Adobe Experience Manager and Target/Analytics, using Java; also managed work of 3 fellow developers.
- Timely resolved high level of customer issues by managing Kubernetes clusters for AEM as cloudservice.
AEM multi-channel and new project ownership
- Led and accomplished smooth transfer of ownership of ’Integrations’ (10 code repositories). Within 5 months of ownership, the product was successfully made cloud-native, leading to the on-boarding of 10+ new customers.
- Contributed to the development of multi-channel interactive forms in AEM Forms by developing the web and print-based channel using Java and JavaScript.
Rest APIs and CCR UI
- Oversaw and managed development of ’Agent UI’ product using AngularJS; improved performance by 15% (for client Deutsche Bank).
- Developed RESTful APIs, UI in AngularJS and hosted the NLP based contextualization engine-Manthan on a multi-cloud container platform-Ethos; product was showcased in Adobe Tech Summit 2017.
UI/UX and NLP based
- Redesigned UI of Hello English app’s home-page and user-profile page using Proto.io. Acknowledged directly by the CEO for the new Material Design based UI.
- Contributed to NLP-based text readability feature: send news articles to users based on their reading abilities.
Education
During my time at NYU I learnt most of my key skills that have I have taken through my career such as teamwork and working to tight deadlines. I grabbed the opportunities to work as a Research Assistant, as a CTO-developer as well as a Grader and a Tutor.
Malaviya National Institute of Technology Jaipur
B.Tech Computer Science
2012 - 2016, GPA 8.52
mnit.ac.in
During my time at MNIT I developed my passion for algorithms and coding. While exploring the vast field of Computer Science I was also actively involved in various campus activities and responsibilities. I lead the Technical Society there in my 3rd year of undergrad and was also part of the college dance group in my early years of college.
Projects
Where do I deploy my particular ML application?
In this project we tried to compare three environments i.e., the bare Virtual Machine, Docker running inside the Virtual Machine, and Singularity running inside the Virtual Machine on two different workloads i.e., an MNIST digit classifier using Convolutional Neural Network and a Sentiment Analyzer using Recurrent Neural Network. We used the same underlying hardware and ran all the workloads using a GPU-enabled Virtual Machine inside the Google Cloud Platform to better compare the metrics and understand the implication of the change of metrics in different environments on the performance of the application.
Song recommendations that are fresh to the ears and fit your listening habits.
It is important for a recommendation system to minimize the user’s effort for providing feedback and to maximize the user’s satisfaction by playing the appropriate song at the right time. In this project, we leveraged the existing user’s song organization data alongwith the track’s metadata to analyze and recommend new tracks to the listener.
A good movie poster can convey a film's theme and genre to a wider audience,
Here we examined not only the visual aspects of the poster but also analyzed the textual information, to see if it can reinforce the model. In combination with textual learning, ResNet50 was able to predict movie poster genres fairly accurately. This suggests that the visual elements included in movie posters of a given genre are relatively consistent.
Data Science can play a big role in getting insights from uncharted territories.
Through this project, we investigated the predictive correlation and causation between (1) Average Daily Traffic of an area and (2) Number of reports of potholes and road damages in the same area.
Quora was popular at that time, so something inspired by that?
It was a team of 2 and we developed a Q&A portal using PHP and MySQL. We were inspired by the success of Quora and got inspired to create something similar to that; received AA in the project.
The then existing mitigation techniques failed to operate when the server is under massive DDoS attack.
The proposed technique ensures the availability of enough resources during massive DDoS attacks for the existing DDoS mitigation techniques. This is done by blocking the attack traffic for certain time and making the resources available to the DDoS mitigation tools.
A Little More About Me
Alongside my interests software engineering some of my other interests and hobbies are:
- Sketching
- Trekking
- Dancing