Sukanya Pal

AI Researcher

Serampore, West Bengal, India3 yrs 9 mos experience
Most Likely To Switch

Key Highlights

  • Reduced loss ratios significantly through innovative ML models.
  • Designed comprehensive data science roadmaps for business solutions.
  • Led teams to identify millions in additional commission opportunities.
Stackforce AI infers this person is a Data Scientist with expertise in InsurTech and advanced machine learning applications.

Contact

Skills

Core Skills

Machine LearningNlpComputer VisionRoute Planning

Other Skills

Cluster AnalysisData AnalysisData ScienceExploratory Data AnalysisJavaLeadershipLogistic RegressionProject ManagementPython (Programming Language)Regression AnalysisSnowflakeStaff SchedulingStatistical Data AnalysisSupervised LearningSwift

About

I am a data scientist with 4 years of industry experience and a Master's degree in Big Data Analytics. I love to explore complex business problems and solve them with data science and machine learning. To accomplish this, I focus on extracting key insights from data, building amazing models to target industry needs, and keeping myself up-to-date with new innovations in the field. Skills: Statistical Analysis | EDA | Visualisation | Regression | Machine Learning | NLP | Deep Learning | Computer Vision Languages/Tools: Python | Pandas | Numpy | Scikit-Learn | Imblearn | SQL | Google OR-tools | Tableau | MS Excel | AWS | GitHub | R Domains: InsurTech | Vehicle analytics | Supply chain management | Healthcare My personal and professionals experiences have taught me the follow about myself: I can independently create, scale and optimise intricate supervised, unsupervised and deep learning models to innovate, reduce loss ratios and improve risk identifications. - I put the project goal at the centre of all that I do - My gut instinct is good but I seek data for my decisions - I experiment with latest industry technologies and methodologies Impacts: - I created supervised models on imbalance claims data to reduce loss ratios from 90% to 38% and trained unsupervised ML models to classify businesses into risk categories that reduces manual labor by 14% - I build mode of travel predictor, consignment route planner and vehicle damage detector engines using vehicle analytics, adding innovative products on company portfolio I can design data science roadmaps from vague business problems to enhance data-centric solutions within the organisation. - I can efficiently align non-technical visions with technical strategies - I excel at the execution of big ideas with given constraints - I am competent in breaking down full scale tasks into step by step deliverables Impact: - I created detailed roadmap to build US businesses feature store using NLP (entity resolution) I master leadership skills that lead to successful teams - I create an environment where everyone can interact freely and contribute - I regularly arrange 1:1 with team members and provide guidance to conquer roadblocks - I promote cross-group collaboration Impact: - The DS team identified ~2MM of additional commission under by project leadership I educate, refine, and drive myself to be a better person: - I am constantly learning because I never settle - I stay calm when faced with adversity - I focus on making high-quality decisions

Experience

Sensight technologies private limited

Technical Leader (Data analytics and AI)

Nov 2023Present · 2 yrs 4 mos · Bengaluru, Karnataka, India · On-site

Zapscale

Data Scientist

Nov 2022Nov 2023 · 1 yr · Kolkata, West Bengal, India · Hybrid

Embroker

Data Scientist

May 2022Oct 2022 · 5 mos · San Francisco Bay Area (Remote)

  • Reduced ESP loss ratio from 90% to 38% by flagging 46% policies through supervised machine learning models built using demographic data.
  • As a team lead and project manager, provided strategic guidance and vision to the team that led to the identification of $1.5 - 1.9 MM in additional commission opportunities over 63 products through optimization of broker-wholesaler economic terms.
  • Designed a detailed roadmap and algorithm for performing NLP based entity resolution to create normalized feature vector for all US businesses.
  • Predicted employee count and funding growth rates for companies and classified them based on riskiness which reduced underwriter manual effort by 14%.
  • Contributed extensively in building a brand new data science and engineering team by reviewing assignments, conducting technical interviews and providing onboading support.
Supervised LearningMachine LearningNLPStatistical Data AnalysisData ScienceProject Management

Sensight technologies private limited

2 roles

Data Scientist

May 2019May 2022 · 3 yrs · Bengaluru, Karnataka, India

  • Route Deviation and familiarity: Designed an algorithm to identify
  • deviations from pre-defined routes in real time using minimal computations to
  • reduce load on smartphone battery. Also devised a scoring technique for route
  • familiarity of a driver.
  • Vehicle Damage detection: Developed an algorithm to verify damage of car
  • and classify location and severity of damage using images by transfer learning which helped in auto-verification of damage during insurance claims.
  • Route planning and staff scheduling: Solved TSP and staff scheduling
  • with real life constraints using google OR-TOOLs. Used clustering techniques to
  • reduce the scale of problem. Also developed a desktop GUI to run the solution.
  • Driver and mode detection: Used ML to predict the mode of travel and also
  • to predict if the user was driving in case the mode was car. This helped in better risk segregation.
  • Scoring using smartphone: Extension of the work in internship. Handled the
  • data challenges due to the orientation independence and mobility of a
  • smartphone. Made suitable changes in the algorithm to implement the same for
  • 2-wheeler drivers. Also translated code from python to java (for android) and
  • swift (for IOS) for easy app development.
Route PlanningMachine LearningComputer VisionStatistical Data Analysis

Data Science Intern

Jan 2019Apr 2019 · 3 mos · Bengaluru, Karnataka, India

  • Automobile driver safety scoring: Identified the key factors of risk while
  • driving and churned the available data from OBD device to create parameters that
  • represent those risks. Also designed an algorithm to score the drivers on a scale of
  • 1 to 100 based on combining the different safety parameters.
  • Fuel mileage prediction: Developed a supervised machine learning based
  • model to predict the fuel mileage using driving habits and patterns.
Machine LearningData Analysis

Education

Ramakrishna Mission Vivekananda Educational and Research Institute

Master's degree — Big Data Analytics

Jun 2017Jul 2019

Serampore College

Bachelor's degree — Physics

Jul 2014Jun 2017

Stackforce found 100+ more professionals with Machine Learning & Nlp

Explore similar profiles based on matching skills and experience