Mehak Goyal

AI Researcher

United States8 yrs 9 mos experience

Most Likely To SwitchHighly Stable

Key Highlights

Expert in Machine Learning and Deep Learning.
Led significant projects at Microsoft improving AI capabilities.
Published research in Amazon ML Conference.

Stackforce AI infers this person is a SaaS-focused AI specialist with strong expertise in machine learning and deep learning.

Contact

Skills

Core Skills

Machine LearningDeep Learning

Other Skills

Reinforcement LearningPythonPytorchC++MXNet

About

In a line - I am learning to make AI products. I am interested in learning everything - non tech things too! Do get in touch if I can be of help. Dropping in a note with a connection request would be super-helpful! :)

Experience

8 yrs 9 mos

Total Experience

2 yrs 11 mos

Average Tenure

5 yrs 3 mos

Current Experience

Microsoft

2 roles

Senior Applied Scientist

Promoted

Jan 2025 – Present · 1 yr 3 mos · Seattle, Washington, United States

Works on enterprise search in Microsoft Copilot & Copilot Search!

Machine LearningDeep LearningReinforcement LearningPythonPytorch

Applied Scientist 2

Jan 2021 – Jan 2025 · 4 yrs · Seattle, Washington, United States

Leading the effort for ranker unification across lexical & semantic relevance rankers. (2025-Present)
Continue working on semantic search based file search (personalization) grounding for RAG based Microsoft copilot. (2024- 2024)
Currently, working on improving the file search (personalization) grounding for RAG based Microsoft copilot. (2023- 2024)
Implemented a graph based document embedding model to improve the related document list recommendation for M365 users. (2022-2023)
Developed a transformer based ranking model to improve the job answer ranking in Bing Search. (2021-2022)

Machine LearningDeep LearningPythonPytorchC++

Amazon

Applied Scientist Intern

Jan 2020 – Jan 2020 · 0 mo

Worked as part of the Contextual AI Stack Team, Alexa AI - Natural Understanding.
Project: Deep Reinforcement Learning for Domain Classifiers.
Developing algorithms for online training of classification models using live traffic.
Researched and implemented policy gradient (REINFORCE algorithm), deep q learning and multi-armed bandit algorithms for a BERT based model.
Published research paper in the Amazon ML Conference.

Deep LearningReinforcement LearningPython

Adobe

Member Of Technical Staff

Jan 2018 – Jan 2019 · 1 yr

Provided fixes for high-risk bugs and customer cases reported for OEM release and internal use.
Tasked with the responsibility to integrate the Adobe PDFL suite in APPE before the product release.