somya singhal

AI Researcher

Noida, Uttar Pradesh, India7 yrs 8 mos experience

Highly Stable

Key Highlights

Expert in building machine learning models.
Proven track record in data-driven decision making.
Strong background in user segmentation and recommendation systems.

Stackforce AI infers this person is a Data Science expert with a focus on machine learning and risk modeling in Fintech and SaaS.

Contact

somyasinghal@fb.com LinkedIn

Skills

Core Skills

Data ScienceMachine LearningProduct Strategy

Other Skills

Bayesian ClassifierCSSClassification ModelsCommunicationCosine SimilarityDashboardsData AnalysisData EngineeringData MatchingHTMLJavaScriptLarge Language ModelsLog ProcessingLogistic RegressionMicrosoft Excel

About

I am IIT Roorkee alumni with masters in data science from Columbia university. I have rich work experience working with meta, Samsung, yahoo, Siemens and Navi.

Experience

Samsung r&d institute india

Senior Staff Data Scientist

Mar 2022 – Jun 2024 · 2 yrs 3 mos · Bengaluru, Karnataka, India · On-site

Built nlp framework which can run multiple large language models to solve multiple customer problems like log processing, text translation, text summarization.

NLPLarge Language ModelsLog ProcessingText TranslationText SummarizationData Science+1

Navi

Lead Data Scientist

Jun 2021 – Feb 2022 · 8 mos · Bengaluru, Karnataka, India · On-site

Building risk models for underwriting engine: XGBoost classification model to identify non-starter fraud people who do not pay any emi. Also, built XGBoost model to help identify risk with the new to credit customers in giving loans.

XGBoostRisk ModelingClassification ModelsData ScienceMachine Learning

Findem

Lead Data Scientist

Mar 2020 – Oct 2020 · 7 mos · San Francisco Bay Area

Machine Learning: Logistic Regression classification model to match LinkedIn data with Github data and StackOverflow data. Compute cosine similarity scores between the profiles on LinkedIn and GitHub and stack overflow using the data - name,location and skills. Feed similarity scores to logistic regression classifier to get the probability of match of profiles.

Logistic RegressionCosine SimilarityData MatchingData ScienceMachine Learning

Facebook

Senior Data Scientist

Jan 2019 – Mar 2020 · 1 yr 2 mos · Menlo Park, California, United States

Infrastructure Strategy: Partnering with the PM and Engineers to inform and support the product decisions by understanding ecosystems, user behaviors, moving trends, and monitoring key product metrics, understanding root causes of changes in metrics. Building data engineering pipelines, data analysis to drive product decisions, instrumenting product metrics and displaying it on dashboard, machine learning projects.

Data EngineeringProduct MetricsMachine LearningData ScienceProduct Strategy

Samsung r&d institute india

Lead Data Scientist

Jul 2014 – Jul 2017 · 3 yrs · Bangalore Urban, Karnataka, India

User Segmentation: Segmented users into various clusters based on the interest profile and developed the Bayesian classifier to classify new user from MyGalaxy to one of the formed clusters.
UserCategorization: SVMclassifiertodohierarchical multilabel classification of apps, improvedusing active learning and user categorization to Fitness and Wellness category using Apache Spark.
User Profile: Developed user profile based on interest, age, location, gender and habit. Contributed in creating interest-based profile which categorizes user into various categories like shopping, toys, shoes.
App Recommendation Engine: Recommendation of apps to user on Samsung MyGalaxy application using a hybrid approach combining popularity of apps and matrix factorization model of collaborative filtering.

User SegmentationBayesian ClassifierSVM ClassifierRecommendation SystemsData ScienceMachine Learning