Help & Supportcareers@gravityer.com
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Staff ML Scientist @ Etsy | Ex-Staff MLE @ Shopify | Founding MLE @ HubSpot AI | CS @ Georgia Tech
www.mukulsurajiwale.com Staff Machine Learning Engineer with 9+ years of experience building and scaling end-to-end machine learning systems in Python, using tools like TensorFlow, PyTorch, Keras, Scikit-learn, and Hugging Face Transformers. I’ve led high-impact projects across the full machine learning lifecycle, from problem framing and data collection to model development, deployment, monitoring, and continuous improvement. I enjoy solving meaningful non-trivial problems at the intersection of research and application and love building things from zero to one with likeminded mission-driven people who care deeply about the societal impact of their work. Currently, I am working at Etsy as a Staff Machine Learning Scientist. Before Etsy, I worked at Writer, where I led the development of a next-generation multimodal, agentic, vision-based information retrieval system to enable accurate natural language QA. Before Writer, I was a Staff Machine Learning Engineer at Shopify, where I led applied research on LLMs and information retrieval for e-commerce. Before Shopify, I was a founding member of the Artificial Intelligence group at HubSpot, where I rose to become the most senior IC MLE and helped grow the team from 5 to XX+ members. My work covers the following: - Training and deploying LLMs for multimodal information retrieval and QA. - Building LLM fine-tuning, evaluation, and serving layers for HubSpot’s ML-Ops platform - Fine-tuning and serving LLM models for specific use case using SFT and DPO - Real-time streaming anomaly detection and forecasting on time series data - LLM-based topic detection in conversations - Bayesian Active Learning - Markov chains for user action prediction - Text classification, clustering, and similarity - Search: semantic retrieval, ranking, and performance monitoring - Contextual multi-armed bandits for adaptive A/B testing - De
Georgia Institute of Technology
Master of Science - MS, Computer Science, Concentration in Machine Learning
N/A – Present
Rensselaer Polytechnic Institute
Bachelor of Science - BS, Computer Science, Minor in Business
N/A – Present
Etsy
Staff Machine Learning Scientist
September 1, 2025 – Present
New York, New York, United States · Hybrid
Writer
Staff Machine Learning Engineer - Multimodality
January 1, 2025 – September 1, 2025
Melodi
Advisor
January 1, 2024 – Present
Shopify
Staff Machine Learning Engineer
January 1, 2024 – January 1, 2025
New York, New York, United States
HubSpot
Senior Machine Learning Engineer II
January 1, 2021 – January 1, 2024
HubSpot
Senior Machine Learning Engineer I
January 1, 2018 – January 1, 2021
HubSpot
Machine Learning Engineer
January 1, 2017 – January 1, 2018
Onshape Inc.
Software Engineering Intern
January 1, 2016 – January 1, 2016
Cambridge, MA
Rensselaer Polytechnic Institute
Lighting Research Center: Undergraduate Researcher
January 1, 2015 – January 1, 2015
Troy, NY
National Instruments
R&D Software Engineering Intern
January 1, 2015 – January 1, 2015
Austin, TX
Monster Worldwide
Software Engineering Intern
January 1, 2014 – January 1, 2014
Weston, MA
Kumon North America, Inc.
Teaching Assistant
February 1, 2011 – June 1, 2013
DeepSelfie
April 1, 2017 – May 1, 2017
- Designed and implemented a deep convolutional neural network model that is able to determine the screen coordinates of where a person is looking on their phone based on a selfie. - Python, TensorFlow
Massively Parallel Deep Neural Network Framework
February 1, 2017 – May 1, 2017
- Used C++ and the OpenMPI message passing framework to design and implement a parallel deep neural network framework capable of performing multiclass classification. - Performance tested various large network architectures on RPI's AMOS supercomputer, which runs on the IBM Blue Gene/Q architecture. - Achieved significant forward and backward propagation execution speedups.
Vivarray
January 1, 2017 – Present
- Developed a comprehensive business model for Vivarray, a virtual reality live streaming service for sporting events. - Aided team in performing extensive industry research and customer discovery including talking with representatives from GoPro and several sporting event venues. - Pitched Vivarray to several venture capitalists, IP lawyers, and successful entrepreneurs. - Placed 3rd out of 36 teams in RPI's annual business model competition.
Commune
October 1, 2016 – November 1, 2016
- Lead developer. - Led a four-person team to develop an iOS and Android app that makes communal living easier. - App is essentially an issue tracker for real life (think JIRA for real life). - App also allows users to keep track of total and individual monetary contributions in a group. - Swift 3.0 and Firebase
SafeView
September 1, 2016 – November 1, 2016
- Team lead. - Led a four-person team through the idea generation, customer discovery, product market fit, and business plan (Business Model Canvas) design processes for our product SafeView. - SafeView is an easy to install wireless camera kit that can prevent you from being one of the half a million Americans that get into car accidents due to improper lane changing every year. - Prepared and delivered several presentations concluding with a final pitch.
Inquiry
November 1, 2015 – Present
- Facebook Global Hackathon 2015 2nd place winner. - Chrome extension that allows you to search within any video on Youtube. - Search results are based on optical character recognition, image classification using machine learning from an external API, and close captioning data. - Python, OpenCV, Scala Check out demo video here! https://youtu.be/gtay0N8pzrA?feature=shared
inCync
September 1, 2015 – Present
- HackMIT 2015 Finalist. - Facebook and General Electric prize winning presentation app. - Users can create a presentation by setting the duration and the time intervals at which they want to be reminded. They can then put their phone in their pocket and it will gently vibrate at the specified intervals thus allowing the user to effortlessly keep track of how much time has passed. - Multiple users can join a presentation. All phones will be in sync and vibrate at the same time. - Integrated with Microsoft Powerpoint to allow users to easily set vibration intervals while making their presentation. - Javascript, NodeJS, Ionic, Parse.
Galleri
October 1, 2014 – Present
- Developed an cloud based iOS app that allows users to take pictures and group them in albums that can be shared. - Multiple users can add pictures to albums that they are invited to. - Xcode, Objective-C, Parse, Sketch 3.
Cultural Fit Analysis
The candidate's diverse project portfolio, including hackathon wins and personal projects, showcases innovation, initiative, and a passion for technology beyond their core job responsibilities. Their experience across various companies (Etsy, Writer, Shopify, HubSpot) and roles, coupled with advisory work, suggests adaptability and a broad understanding of different organizational contexts. The focus on cutting-edge ML technologies and leadership roles aligns well with a dynamic, growth-oriented culture.
Soft Skills & Operational Fit
The candidate demonstrates strong leadership through team lead roles in projects and professional experience (e.g., 'Led a four-person team', 'Staff MLE on the team, I helped set the technical roadmap, mentored'). Their involvement in shaping ML hiring and assessment processes at HubSpot suggests a proactive and collaborative approach to team and organizational development. The detailed project descriptions and professional experience indicate strong communication skills and an ability to articulate complex technical work.