Conversational AI platform built at Haptik has been powering chatbots since 2013, way before chatbots were cool. We are currently looking to grow our Machine Learning team to strengthen and expand multiple research projects in the domain of information retrieval, intent and entity extraction from complex dialogues, personalising chatbots with emotions and personality. As Machine Learning Engineer, you should be willing to deep dive into Haptik’s exponentially growing text data consisting of more than 2 billion messages exchanged over more than 300 enterprise bots and build scalable, personalized and robust dialogue system using Natural Language Processing and deep learning.
Understand various components of existing Machine Learning Pipeline at Haptik
Contribute to research team during experimentation phase and help them narrow down in conclusive direction
Experiments and prototype new ideas in the domain of transfer learning to better leverage data sources across languages using NLP and deep learning algorithms
Benchmark the prototype on Haptik’s Data to gain required precision
Perform in depth analysis of failure points of algorithm on different environments
Improve precision and recall of existing modules by using data generated by human agents at the failure point of current algorithm.
Scale models for multiple domains with help of devops, backend and frontend teams.
Foresee data requirements for future research projects and come up with innovative approaches for data generation
Publish and present successful research in relevant journals and conferences
3+ years of industry experience or Post graduation with 2+ years of research experience in Machine Learning projects
In depth experience working on at-least one dataset out of natural language text, images, audio clips, system logs or clickstreams.
Hands on experience on at-least one of the projects amongst information retrieval, machine comprehension, entity recognition, intent detection, semantic frame parsing, machine translation or speech recognition
Experience working with SQL, NoSQL or graph Database
Proficient at programming in any one language and familiarity with at-least one deep learning framework
Relevant publications, patents or talks delivered in reputed conferences is a bonus
Experience with training and tuning language model on large datasets is a bonus
Experience working on Indian language is bonus
Knowledge of parallel computing and hands on experience with Mapreduce algorithms is a bonus