May 2018 – Apr 2019 · 11 mos · San Francisco Bay Area

• Level AI is a San Francisco based AI startup in the conversation AI space.
• I was the first engineering hire, tasked with building product architecture, backend services and building a team.
• Architected and delivered a conversational AI prototype for customer service reps using Google STT and RASA.
• Designed and implemented the intent/entity extraction layer using RASA NLU trained on synthetic retail dialogue data for different brands or use-cases.

Jan 2016 – Apr 2018 · 2 yrs 3 mos · Greater Delhi Area

• Acadview was an EdTech platform for skilling engineering students on programming for various platforms such as python, web development, machine learning and android. We conducted online and offline lectures to skill over 5000 students across North India. The startup was acquired by Upgrad (a leading edTech company in India in 2018).
• My role at acadview was that of a full stack entrepreneur:
• 1. Built a high performant engineering team from scratch on a monthly budget of $3000, I achieved this by leveraging good folks who needed just the right amount of mentorship on engineering and technology.
• 2. Built a data ingestion framework which could handle excel sheets of different formats and inconsistencies (we were getting semi-structured data from universities). Built a PDF generation pipeline which could scale to millions of PDFs.
• 3. Built the first few version of the LMS (both frontend and backend) that was used to conduct classes, schedule assignments, conduct project evaluations.
• 4. Lead the operations team in both a strategic and executive capacity to select high quality instructors and support.
• 5. Led the content effort to generate relevant content which was in tune with our target audience by using lay-person language, feynman technique, memes.
• 6. Led the product effort to product high quality PRDs for the engineering team.

Oct 2014 – Apr 2015 · 6 mos

• Second engineering hire at frankly. Frankly was a video Q & A platform for celebrities and influencers. One could ask a question to a celebrity and get a video response. In a short time span we on-boarded esteemed personalities from politics, bollywood and sports.
• a. Built the search system using ElasticSearch, the search was done on usernames, query phrases, video meta data and personalized based on user preferences and geography.
• b. Developed a unified notification service (web + mobile) that improved user engagement by 15% and 7-day retention by 20%. Focused on reliable fanout and device targeting.
• c. Picked up a bit of android development as well, built a couple of small features and bug fixes on the app.

Nov 2012 – Sep 2014 · 1 yr 10 mos · Greater Delhi Area

• Quizot was a content platform meant for quizzers and trivia lovers. We differentiated ourselves by carrying highly curated content and targeting a niche category early on. To seed the platform with content I launched a virtual internship program. At it’s peak the platform had ~30,000 MAU with 800,000 pieces of content consumed every month. We were bootstrapped.
• I learnt a lot about software development, product development, user delight via engineering, content management, user generated content, product management during this stint.

May 2011 – Nov 2012 · 1 yr 6 mos · Hyderabad

• Started my career as a QA (manual testing)