Nishanth Chandrasekar

Experience

• 

  Vision Research Foundation

  May 2013 - Jul 2013

  Summer Research Intern
• 

  National Brain Research Centre

  May 2014 - Jul 2014

  Summer Research Intern
• 

  Indian Institute of Science (IISc)

  Apr 2016 - Sept 2018

  Research Assistant

  Worked under Dr. Supratim Ray at the Center for Neuroscience, IISc. Project was mainly on studying Gamma oscillations in the V1 area of the brain - developing nonlinear models as well as analyzing data using signal processing.

• 

  Independent

  Feb 2019 - Jun 2020

  Machine Learning Engineer

  Experience involving Computer Vision - have built production level systems to automate processes on the cloud. Some highlights - 1. AIPARAS - Object detection, Classification - automatically detecting and assessing advertising displays in Indian kirana stores for adherence to marketing norms in low light/bad photo quality conditions. Based on object detection and ensembling.2. Error Correction using transformers - Using NER + Text Classification with transformers to efficiently preprocess data and reduce manual error detection work, also using seq2seq models for automatic error correction. Show less

• 

  Picquora.ai

  Jun 2020 - Jul 2021

  Machine Learning Engineer

  Have worked on various AI and Deep learning techniques for solving varieties of problems including Computer Vision, NLP as well as problems involving multi modal data. Experience with a variety of models from domains including CNNs, transformers as well as combinations of the two.

• 

  Salesken

  Jul 2021 - Jul 2022

  Senior Machine Learning Engineer

  The product aimed to increase sales agent conversion rate and to provide valuable insights to sales agents and managers on their sales pipeline. This was done by providing both real time and post call insights.Conversational AI, NLPCuing: Providing real time cues to sales agents by fetching relevant product specific information in response to customer inputs. Data exploration, System Design and development.The experimentation spawned multiple side projects that eventually went into the product.Challenges included - designing for self serve and to work in real time.Methods - Information Retrieval, Retrieve and re-rank, Search, web scraping, document parsing, transfer learningQuestion Detection:A question detection AI model adapted to conversation data, later combined with various independent models to play a role in two components of the product – Consultative Selling and Call Summarization Experimented with knowledge distillation and model optimization to greatly increase throughput and reduce latency with minimal loss of performance Consultative Selling: Data analysis to combine the question detection model with other AI models to provide a metric for a sales agent’s consultative selling ability that correlates with successful conversionCall Summary: Developing a module that provided post call insights to any sales agent/manager that was making a follow up call to a customer. Relevant points of interest from previous calls to the customer had to be extracted and displayed in a crisp manner to catch the agent up on the sales situationModule formulation, System design, development and deploymentProject Lead - coordinating between and brainstorming with developers and communicating with data labeling teams on best practices to reduce subjectivity and labeling errorNER, Text Classification, Semantic Similarity Show less

• 

  Mesh

  Sept 2022 - Oct 2024

  Senior Machine Learning Engineer
• 

  Pageloop

  Oct 2024 - now

  Co-Founder