“I met Krupal in 2014 when was a kid just 2 years out of college. Much like any 20-something old, he had many dreams and aspirations. But what impressed me most was that unlike many of his peers, he was willing to jump and take a chance on a relatively unknown startup called Haptik. He joined us as the 1st machine learning engineer, the only person who could speak Data Science in the company. From there on, he built the platform from 0 to 4 BILLION interactions processed. Krupal was the closest it came to the "heart" of the company after the 2 Founders. He was the engine behind the scenes taking Haptik from a startup to a $100 million company. His biggest strength is to not get excited by the technology, but rather the end impact it can have. He built, grew and mentored a team of stellar data scientists who all looked up to him. Krupal will go on to do amazing things in life and I hope to work with him again during that course.”
About
Experience & Education
-
Amazon
View Krupal’s full experience
See their title, tenure and more.
or
Licenses & Certifications
Publications
-
HINT3: Raising the bar for Intent Detection in the Wild
EMNLP-2020 - Insights from Negative Results in NLP
See publicationIntent Detection systems in the real world are exposed to complexities of imbalanced datasets containing varying perception of intent, unintended correlations and domain-specific aberrations. To facilitate benchmarking which can reflect near real-world scenarios, we introduce 3 new datasets created from live chatbots in diverse domains. Unlike most existing datasets that are crowdsourced, our datasets contain real user queries received by the chatbots and facilitates penalising unwanted…
Intent Detection systems in the real world are exposed to complexities of imbalanced datasets containing varying perception of intent, unintended correlations and domain-specific aberrations. To facilitate benchmarking which can reflect near real-world scenarios, we introduce 3 new datasets created from live chatbots in diverse domains. Unlike most existing datasets that are crowdsourced, our datasets contain real user queries received by the chatbots and facilitates penalising unwanted correlations grasped during the training process. We evaluate 4 NLU platforms and a BERT based classifier and find that performance saturates at inadequate levels on test sets because all systems latch on to unintended patterns in training data.
-
Production Ready Chatbots: Generate if not Retrieve
Workshop on Reasoning and Learning for Human-Machine Dialogues, AAAI 2018, New Orleans, LA, USA
See publicationIn this paper, we present a hybrid model that combines a neural conversational model and a rule-based graph dialogue system that assists users in scheduling reminders through a chat conversation. The graph based system has high precision and provides a grammatically accurate response but has a low recall. The neural conversation model can cater to a variety of requests, as it generates the responses word by word as opposed to using canned responses. The hybrid system shows significant…
In this paper, we present a hybrid model that combines a neural conversational model and a rule-based graph dialogue system that assists users in scheduling reminders through a chat conversation. The graph based system has high precision and provides a grammatically accurate response but has a low recall. The neural conversation model can cater to a variety of requests, as it generates the responses word by word as opposed to using canned responses. The hybrid system shows significant improvements over the existing baseline system of rule based approach and caters to complex queries with a domain-restricted neural model. Restricting the conversation topic and combination of graph based retrieval system with a neural generative model makes the final system robust enough for a real world application.
Patents
Recommendations received
3 people have recommended Krupal
Join now to viewOther similar profiles
Explore more posts
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More