Consulting startups on how to use AI and Data to make better products for their end customers

Pixanimate harnesses the power of Generative AI to redefine video creation. Our platform facilitates the generation of high-quality, customizable videos with just a click of a button.We're equipped for fast turnaround, ensuring your message reaches your audience without delay. Our diverse template options and affordable pricing plans make our services accessible to businesses of all sizes.Our cutting-edge technology includes integration with Chat GPT for generating natural language voice-overs, and a seamless API for integrating our AI video creation technology into your existing platforms. From translating complex information into easily digestible videos to producing captivating eCommerce content, Pixanimate leverages Generative AI to take video content to the next level.www.pixanimate.com

Responsible for delivering end to end data needs of the product from all fronts - platform, engineering, analytics and data science

Working on Personalisation, relevance and ranking problems to increase efficiency/CTRs in the customer journey funnel to increase overall conversion of the platform at scale of swiggy (~40+ Mn customers). Not just the model building, even deploying them at this scale and making sure things are working as we want by collaborating with product, engineering and analytics.More precisely problems like-1. Restaurant ranking2. Menu personalisation3. Payment gateway routing4. Preferred payment option selection5. Cross-sell stuff

I worked with roamanalytics.com here.- The main work was around building a knowledge graph around US healthcare domain.- Work involved a lot of NLP which includes getting out Named Entities (using semantics, syntactics and ML) from unstructured text and populating that in the Neo4j Graph. Also used lot of ML to make connections in the graph to make it interesting and informative.- Tool-wise: Python left, right center, all DS libraries Pandas, numpy, sci-kit, Spark, Neo4j , Airflow, AWS infrastructure etc.

Lendingkart is a Fintech operating in B2B domain.Summary of my stay there in terms of projects and skills I acquired.1. Designed the entire mobile analytics platform and coded it in python. (Learned a lot of software engineering stuff: AWS, Testing, Deployment etc)2. Analysing the user SMSs for getting insight into her credit rating. Learnt a lot of NLP during this which includes:a) Semantic and Syntactic Analysis of SMSs.b) Training word vectors for the financial SMS context and tried to use them for certain tasks.c) Clustering of documents to get more insight into user behaviour.

Story-It was an early stage startup with one main motive- growth. Team building was in progress and there was minimal interaction between teams. I with one or two more people connected those dots. As I knew Technology as well as had a good understanding of business, started exploring business and Technology from all aspects. Interacted with different teams and made them realise the relevance of other departments and how they can use the work done by others. Being in a startup you have to do everything which comes your way. I also had to go through this interesting path- I've done sales, marketing, operations, data stuff, giving training, motivating people etc etc...If I've to put it formally, then- - Explored many important KPIs which made business more quantitative and objective.- Defined many critical processes which people use everyday.- Made many interactive and real time dashboards for personal use, and later after many iterations passed it to entire operations team across cities.- Designed, monitored and analysed ROIs of various promotional offers and did predictive analysis for exploring more into the schemes.- Modelled the cluster business model using machine learning which made it more dynamic and most importantly self scalable as new clusters add. But due to business constraints couldn't implement it at that stage.- Did proper segmentation of clients and many other stakeholders which helped us target the correct audience.