Nov 2024 – Present · 1 yr 4 mos · Gurugram, Haryana, India

Sep 2024 – Apr 2025 · 7 mos · Remote

Sep 2020 – Aug 2021 · 11 mos · Gurugram, Haryana, India · On-site

• Built YodaApp - Instagram for learning where you consume knowledge via user-generated bite-sized infographics. Scaled to 500K users and 10K creators. Raised $1M but failed to achieve PMF hence returned the capital back.

Jul 2019 – Dec 2019 · 5 mos · Gurugram, Haryana, India

• Gurgaon Orators Toastmaster Club is one of the best toastmasters club in Gurgaon. Toastmasters International is a non-profit educational organization that teaches public speaking and leadership skills through a worldwide network of clubs. Headquartered in Englewood, Colo., the organization's membership exceeds 357,000 in more than 16,600 clubs in 143 countries. Since 1924, Toastmasters International has helped people from diverse backgrounds become more confident speakers, communicators, and leaders.

Jan 2019 – Jun 2019 · 5 mos · Gurugram, Haryana, India

• Gurgaon Orators Toastmaster Club is one of the best toastmasters club in Gurgaon. Toastmasters International is a non-profit educational organization that teaches public speaking and leadership skills through a worldwide network of clubs. Headquartered in Englewood, Colo., the organization's membership exceeds 357,000 in more than 16,600 clubs in 143 countries. Since 1924, Toastmasters International has helped people from diverse backgrounds become more confident speakers, communicators, and leaders.

Apr 2016 – Oct 2024 · 8 yrs 6 mos · Gurugram, Haryana, India

• Built Trainman from 1st line of code to the most feature-rich train app of the country. Scaled it to 25M users. Started with zero and learnt everything on the way.
• Tech: Build entire Trainman stack including India's first predictive AI model for train travellers which tells the confirmation likelihood. Added a guarantee to it by launching Trip Assurance - Free upgrade to flight if train doesn't get confirmed. Built AI, Backend, DevOps for the product.
• Product: Launched lot of industry-first product in train space - PNR prediction, Past trends, Refund Schedule, Seat Availability Calendar, Trip Assurance & a lot more.
• Growth: Scaled the traffic from zero to 25M users and to a MAV of 10M - purely organic with close to zero marketing budget.
• People: Started with 2 and eventually became 30 members strong until acquisition. Each one of these 30 significantly contributed to the Trainman growth story.
• Business: Bootstrapped it to profitability & then raised money to further accelerate growth. Sold it to the Adani Group.
• Awards & Recognitions: Trainman was selected in Google For Startups Accelerator program (Class 7)
• Investors happy, Employees happy, Founders happy!

Sep 2014 – Feb 2016 · 1 yr 5 mos · Gurugram, Haryana, India

• Mailing system with email masking functionality: Built a full-fledged email server in python-django using postfix which is capable of sending and receiving emails. It can also automatically forward emails based on certain rules which enables the masking of actual email address of sender/recipients from one another. Used message queues (rabbitmq) to decrease response time and increase scalibility. It was easily able to handle more than 500 emails every minute. Same is used for all communications between sellers and customers on weareholidays platform.
• Seller-customer forum: Built a forum in python which provided all functionalities like sending message and/or attachments, changing subject, adding/removing members etc. In second phase, forum was used instead of mailing server for all interaction between sellers, customers and account managers.
• Weareholidays marketplace for booking packages: Weareholidays marketplace enables seller to look & cater to relevant customer queries for travel package booking. I was involved in building the core pick logic in Java which ensured fair chances of query picking to all the sellers serving the given destination.
• Bia app backend: Led a team of 3 people to build full backend for weareholidays android app "bia" in Java. The app helps users to create beautiful trip journals effortlessly. The functionalities it provides is - User signup using email/facebook, Adding photos/checkins/notes while on trip, capturing routepoints automatically to plot complete trip on map, publish and share trip to make visible for other users, view all the public trips.
• Reporting tool: Designed and developed an internal reporing tool in python-django for generating nearly ten different types of reports on email exchanges. The most common report generated was - seller-wise report which captures all seller details such as mean response time for query, query pick ratio, queries converted to leads-picked ratio among others.

Dec 2011 – Sep 2014 · 2 yrs 9 mos · Gurgaon

• Designed & Developed prototypes & ensured their timely evolution into products:
• Query Handling for semantic search: Worked on Natural Language Processing using python-nltk
• to handle wide range of travel queries (such as transport, places, weather, hotels/restaurants etc) on ixigo trip-planner. Also added location based biasing so that places which are near to user are given more weightage. Developed an automated test-suite in django for query benchmarking.
• Multimodal Transport Engine: Built multimodal transport engine (in java using neo4j graph
• database), which takes into consideration all the common travel mediums (flight, train, bus and taxi) and gives best possible combination to user to go from one city to any other city in the world. The route may contain zero or more hops.
• Generic Sms parser: Developed a generic sms parser in python which can extract meaningful
• information from wide variety of smses of flight and train journeys. This parser is the core of ixigo pnr app (having more than 1 lakh downloads on play store) wherein passenger details are extracted
• automatically from sms.
• Hotels List Sorting: Built a hotel sorting algorithm which uses bias of distance, price and trust-you
• score to show the most relevant hotels to the user first. This algorithm is being used in ixigo brand app.
• Named Entity Detection: Created a named entity recognizer which extracts named entities
• (destinations, places of interest) from various sites (such as wikipedia, wikitravel, mustseeindia etc) and converts them into hyperlinks using information in database.

May 2010 – Dec 2011 · 1 yr 7 mos

• Streamlined operations as a part of Global Database support team and helped remodel GTI towards a service-oriented architecture

May 2008 – Jul 2008 · 2 mos · Vienna, Austria

• Security enhancement in Moodle - an e-learning application