Prashant Anand

Prashant Anand

Machine Learning Engineer




I'm a senior Machine Learning Engineer with 4 years of experience in building scalable, high-performance production ML systems to support a wide range of business needs. Whether it's data analysis, ideation, and experimentation or deployment and maintenance, I'm actively engaged in all aspects of developing ML systems. I have in-depth experience in Machine Learning, with a particular emphasis on Natural Language Processing (NLP) and Large Language Models (LLMs).

I currently work in the Customer Support domain and lead the exploration and application of ML, NLP, and LLMs in this domain. My day-to-day responsibilities include (but are not limited to) the following:

  • Collaborate with stakeholders & TPMs and analyze data to develop hypotheses to solve business problems and validate them through A/B tests.
  • Frame business problems as ML problems and create suitable metrics for ML models and business problems.
  • Build PoCs and prototypes to validate the technical feasibility of the new features and ideas.
  • Create design docs for architecture and technical decisions and guide the technical implementation.
  • Collaborate with the platform and data platform team to set up and maintain the data pipelines that satisfy our team's evolving needs.
  • Build and deploy production-grade microservices in Python and Go with suitable capacity planning, logging, error handling, distributed tracing, monitors with actionable alerts, and auto-scaling.
  • Write design docs for running A/B tests and perform post-test analyses to determine the impact of ML features on the business metrics.
  • Maintenance, enhancement, and retraining of ML models for features running in production.
  • Responsible for incident handling and included in the on-call rotation of the ML and backend microservices owned by my team.
  • Evaluate technical assignments and conduct interviews for hiring mid-career, new grads, and intern ML engineers.

Tech stack that I use for carrying out my day-to-day responsibilities:

  • Data analysis: SQL, BigQuery
  • ML experimentation and model training: Jupyter Notebooks, PyTorch, Hugging Face Transformers, Azure OpenAI, Kubeflow pipelines, MLFlow
  • Model deployment: TorchServe, Kubernetes
  • Microservice development: Python, Go, gRPC, Datadog, Pagerduty, Sentry, Spinnaker, Docker
Work Experience

Work Experience

  • Senior Machine Learning Engineer (MG3), Mercari, Inc.

    Jul, 2021 - Present

    • Implementing a chatbot to answer customer inquiries using a Retrieval Augmented Generation (RAG)-based approach with large language models (LLMs) like GPT.
    • Designed and implemented a routing algorithm for customer inquiries using ML and NLP to reduce the number of transfers between CS agents. This increased the accuracy of assigning skills to inquiries by 8.2% and reduced the average number of transfers per reply by 26.4%.
    • Developed and deployed a template suggestion feature that allows CS agents to quickly select the suitable template for replying to customer inquiries. This reduced the average time spent responding to an inquiry by 2.2% and is used for 31% of customer inquiries by the CS agents.
    • Collaborating with the hiring team and have conducted 12 technical interviews and evaluated 26 technical assignments for hiring mid-career, new grads, and intern ML engineers so far.
    • Created a technical assignment currently in use for company-wide hiring of ML engineers specializing in NLP.
  • Machine Learning Engineer (MG1/MG2), Mercari, Inc.

    Oct, 2019 - Jun, 20211 year 9 months

    • Implemented A/B test framework in the customer support tool used at Mercari by the CS agents to reply to customer inquiries. This allowed us to A/B test new features, including ML ones, understand their impact on business metrics, and make iterative improvements in a data-driven way.
    • Developed a microservice for sending automated replies to specific types of customer inquiries that don't require human intervention using ready-made templates.


  • Programming Languages



  • ML Libraries & Tools

    Azure OpenAI

    Hugging Face Transformers

    Hugging Face Datasets






  • Backend & Devops




    Google Cloud Platform / GCP




    Apache Airflow


    Plotly Dash



  • Electrical Engineering, Bachelor of Technology (B.Tech.), Indian Institute of Technology, Delhi

    Jul, 2015 - Apr, 2019



  • Few Shot Speaker Recognition using Deep Neural Networks , arXiv

    Published on: Apr 17, 2019

    • Developed a few shot speaker identification framework using deep convolutional neural networks with prototypical loss.
    • Performed speaker identification and few shot speaker identification tasks on Voxceleb dataset using Capsule Network, VGG and ResNet34 architectures.
    • Showed generalization capability of the networks on both tasks by performing experiments on VCTK Corpus
  • Data Driven Sensing for Action Recognition using Deep Convolutional Neural Networks , Lecture Notes in Computer Science, vol 11941. Springer, Cham

    Published on: Dec 17, 2019

    • Developed a novel data-driven under-sampling method using sub-pixel convolutional layers and integrated it with Inflated 3D ConvNet for action recognition.
    • Successfully performed action recognition on both UCF-101 and HMDB-51 datasets at multiple (including very high) under-sampling ratios with small drop in accuracy.
  • Compressive Sensing Based Privacy for Fall Detection , LNCS/CCIS, Springer

    Published on: Dec 31, 2019

    • Developed a privacy preserving fall detection framework based on block based compressive sensing and deep learning which works with wide variety of sensing matrices.
  • Artificial Neural Network based Controller Design for SMPS , IEEE

    Published on: Oct 10, 2019

    • Designed a controller using neural network for half-bridge converter based SMPS to replace conventional PID controllers.


  • Photography

  • Reading

  • Coffee

    Hand DripFrench Press
  • Music

    PianoWestern ClassicalMetalProg MetalClassic Rock