Over the last four years at Databricks, my mission was clear: unblock customers and get them to the "aha moment" fast. I have a get-shit-done attitude and believe in actions, not words. Here’s what I delivered:Cloud Deployment Ease: Streamlined deployment for new accounts, enabling users to hit their first "print hello world" on AWS, Azure, and GCP with zero friction.AWS Marketplace Integration: Led the integration of Databricks into the AWS Marketplace, simplifying acquisition and expanding reach in collaboration with AWS.Databricks SaaS: Bootstrapped a self-service SaaS model, offering a simple email-only signup that requires zero cloud admin skills, set to drive customer growth.Customer Insight & Growth: Developed data-driven insights to pinpoint key 'aha moments' and understand growth drivers for users and revenue.I make complex tech accessible and actionable. No fluff.

Noether Technologies helps companies scale up their data infrastructure, and teams focused on data science, machine learning, and product growth.

I work closely with the executive team to discover how the existing Datacoral infrastructure can be adapted and/or modified to drive business growth for their clients.

I built the Mountain View growth team focussed on new customer signup and onboarding experience from scratch. It was a data informed team with an emphasis on experimentation and causal inference.My core responsibilities included hiring, building a data driven culture, leading efforts to identify early indicators for activation (the aha moment), partnering with the executive team to identify shared OKRs, and partnering with PMs, designers, engineers, and analyst to build an onboarding playbook.

I work closely with the executive team and help them in the following areas.1. Data science and machine learning -- building prototypes to drive business growth, interviewing, and hiring.2. Analytics and product growth -- what kind of metrics the company should be looking at to quantify effectiveness of various operations.

My day to day role at Woven involved working closely with the CEO to generate product insights with data. This was also my first time working with Google Cloud Platform, and I loved it!

I lead various student initiatives at Edmodo with the goal of making learning fun and rewarding. My team recently launched Edmodo Play which is our first by students for students initiative. Within a few weeks of its launch, we had more than a million students answering and helping each other out with middle school math questions. Since then we have expanded to other subjects with a rich collection of student generated content. This work has been the most rewarding so far. I see a teacher in every student and there is nothing more exciting than sparking the curiosity of a student and encouraging them to explain it to other students. This is the future of education!Last year I was the main lead for Edmodo Spotlight which is our teacher content platform. The goal was to surface the right content at the right time to a busy teacher. The work was recognized by the Department of Education and I represented Edmodo at the White House in October 2015. I was also part of a panel on open educational resources along with Microsoft and Amazon at the ASU-GSV conference in April 2016.

In this role, I'm responsible for advising the team on-- Bringing in best practices around data tracking and data architecture-- Identifying opportunities for growth using data and user research

In this role, I'm responsible for-- Identifying and benchmarking main conversion funnels -- Bringing in best practices around experiments and growth frameworks-- Collaborate with relevant product and functional partners to deliver a roadmap of key opportunities

Mentoring students and helping them connect the dots between data science and business.

I sunsetted this app in Fall 2016. It was a fun ride!The goal is the build the most trusted dating platform for professionals. Using social signals from Facebook and professional signals from LinkedIn, Cup of Tea curates and matches each user with a date based on the user preferences (age, gender, distance, interests and hobbies, companies and products they like, musical interests etc) and how similarly they play a game within the app.Do reach out if you are interested in working or collaborating with me on this.

Helping nousDECOR build a great data driven product. And in the process, discovering new things and learning new skills.

My main deliverable was to bring in the best industry practices to visualize core business metrics to maximize efficiency of consumption. The lessons I learned leading the "State of the Union" effort at LinkedIn played a significant role here. Unlike many other consulting roles this particular one was fruitful because I had a partner within Shopkick who was just as passionate as I was to make an impact by measuring the right kind of data, consuming it and using it to make data driven decisions.

I spent a lovely one year at Edmodo, worked on a product I really cared about with extreme passion, made amazing friends and learnt to take a product from conception to the launch phase. I am extremely grateful for my time at Edmodo and here are some highlights (not in any particular order of relevance).1. Attracting and hiring talent: I was at the forefront of hiring for Edmodo after I started there. I brought in two experienced engineers (one for recommendation and one for data pipeline and ETL) and one junior engineer to fulfill team needs. I also actively participated in interviewing Data Engineers and Product Managers for various roles within Edmodo.2. Led the effort to thoroughly analyze engagement of Edmodo members to generate series D pitch deck for the CEO. Asked important questions and answered them. Some examples include:(a) What are teachers and students mostly doing on Edmodo and what drives them to these activities?(b) What is the value of Edmodo during and outside school hours?(c) What kinds of teachers are present on Edmodo and what differentiates them? Are they growing? Are they moving from one bucket to another? 3. Planned, executed and delivered the first child centric marketplace with limited resources (typical in a startup) in less than 3 months. The main motivation was my early realization that Teachers are busy and it doesn’t help them if we show 200 search results for “mitosis”. See this SRI report http://www.sri.com/work/publications/research-use-khan-academy-schools-research-brief which validated my initial intuition in identify a severe lack of socially curated content platform as the top concern for busy teachers. When I left Edmodo the product was on internal QA and Production servers.4. To bring in a culture of happy hour I brought and installed the first beer keg within Edmodo. Getting a buy-in from my manager was quite easy :)

In short, data analytics, visualization and automation. While at LinkedIn I successfully led multiple efforts to not only generate business insights but also to democratize data within LinkedIn. Some concrete projects which I led are-- Complete overhaul of internal business insights reporting structure which not only reduced the production time from order of weeks to days but also helped drive the discussions at the executive level meeting in the right direction. The outcome of this project continues to be consumed by the CEO of LinkedIn and by the SVP of Products and User Experience on a weekly basis. The tremendous success of this project led to my promotion at LinkedIn in two quarters.-- An internal tool which helps one track the quality of data and instrumentation as and when products are being launched at LinkedIn. -- Network flow analysis of user behavior pattern on LinkedIn. The overall goal was to quantify how user behavior changes with new product launches. For example, is there an unwanted disruption leading to churn? Are there any products which get cannibalized because of a new product? Do different segments of LinkedIn members use and adapt to products differently?-- Topic modeling for LinkedIn ads using Latent Dirichlet Allocation techniques for better personalization of ads. This work demonstrated a CTR boost of about 2% in offline testing. My work involved analyzing large data sets across multiple sources (Hadoop, SQL, CSV, http://www.bls.gov/ to name a few), clustering member base and understanding how different member segments use LinkedIn and quantifying the value of member actions. These insights are then used to further make LinkedIn a better place for its members.The tools I use routinely include(a) Pig, SQL, Hive, R , Python, Bash/Awk/Sed scripting for multi source data gathering, analytics and automation.(b) Javascript, jQuery, GoogleCharts and Tableau for a rich and interactive result visualization.

Researcher in modeling and control, probabilistic computation and large data set analysis. Development of algorithms for sequential model learning with performance guarantees and constrained memory requirements.--- Developed probabilistic techniques for efficient multiple target data association and tracking to speed up random walk based sampling on graphs by over 7 times compared to state-of-the-art technique in the research community. The main challenge was to reformulate the target tracking problem to take advantage of recent developments in the field of probabilistic graph sparsification. Using a smart choice of variables, I demonstrated that one can approximately solve the original (NP-hard) problem with excellent error control. Probabilistic techniques were develop ed and used to reduce the computational complexity of the system. ­ --- Modeling and developing metrics to quantify and reduce complexity of certain aerospace systems. The main challenge here is to model various aspects of dynamic complexity based on the system dynamics model and transient behavior which has a direct correlation to developmental cost to the industry.

Advanced geometric control, robust area surveillance using multiple autonomous vehicles and Bayesian learning.

Developed fluid mechanics based model to understand locomotion in aquatic species which do not generate vortices but which can move around by changing their shape. In differential geometric terms, the aquatic species starts with a certain shape and comes back to the same shape, i.e., a makes a loop in the shape space. By doing so, the non zero Riemannian curvature of the system guarantees a nontrivial translation in the group (or vertical) direction. See http://goo.gl/2bnVj for the relevant publication in Journal of Fluid Mechanics.

Research in Geometric Control Theory. The main focus was to develope differential geometric tools to stabilize and synchronize a group of robots/vehicles with large region of attraction despite underactuation. Coordination of a network of satellites in space for interferometry purposes is a candidate example. Energy shaping techniques were used to prove convergence and stability.