Mohit Bansal
11.5+ years of experience working as an independent consultant providing services into Data Science, Product pricing and Engineering. Worked in a technical lead cum project management role, managing 8-10 resources in my previous roles.
- My work primarily involves Product pricing, developing business solutions using AI/ML models and analytical dashboards in Power BI/Tableau/Looker.
- 8+ years of management Initiation, Planning, Execution, Monitoring and Closure.
- Project management through Agile methodology for smooth execution and business communication for the projects status and timelines.
- Work on design and develop the roadmap for the organizations data flow and architecture with data pipelines.
- Client management in both offshore and onshore locations with 2.5+ years of onshore exp. working at client locations in UK, Europe, Turkey and UAE.
Technical Expertise:
ML ALGORITHMS: NLP, Text Classification, Linear & Logistic Regression, KNN, Decision Trees, Random Forests, Clustering (K-means), Naive Bayes Theorem, Principal Component Analysis, Market Basket Analysis, LSTM/FBProphet, ARIMA, GBM, Recommendation System etc.
SOFTWARE SKILLS: Python, SQL, Databricks, AWS, GCP, Advanced SAS (Base SAS Certified), VBA, R, Tableau, Power BI, EMBLEM, RADAR & MS Office, Azure Cloud.
DATA OPERATIONS: Data Architecture, Azure Data Factory, AWS Model deployment, Databricks, ML Flow and MLOPS
- Role
Staff Product Data Scientist
- Years of Experience
11.5 years
Skillsets
- AWS - 3 Years
- NLP - 3 Years
- PowerBI - 6 Years
- Computer Vision - 2 Years
- ARIMA - 2 Years
- GBM - 3 Years
- Databricks - 2 Years
- SQL - 8 Years
- Python - 5 Years
- LLMs - 1 Years
- Tableau - 5 Years
Vetted For
- Roles & Skills
- Results
- Details
Staff, Product Data ScientistAI Screening- 59%
- Skills assessed :Analytics, Looker, Communication, Agile, Python, Scrum, SQL, Tableau
- Score: 59/100
Professional Summary
- Apr, 2023 - Present2 yr 4 months
Independent Consultant
Contractual Engagements - Dec, 2021 - Jan, 20231 yr 1 month
Data Architect - ML/Technical Lead
Koantek AI Cloud Pvt. Ltd - Nov, 2019 - Nov, 20212 yr
Manager - Data Science and Analytics
PWC India - Aug, 2012 - Feb, 20152 yr 6 months
Data Analytics Assistant
RSA Insurance - Mar, 2015 - Dec, 20161 yr 9 months
Senior Consultant-Data Analytics
Allianz Insurance - Jan, 2017 - Nov, 20192 yr 10 months
Actuarial Consultant
AXA Insurance
Applications & Tools Known
Python
NLP
SQL
SAS
RStudio
Microsoft Power BI
Tableau CRM
Looker
Oracle SQL Developer
Work History
Independent Consultant
Contractual EngagementsActuarial Consultant/Manager, AInsurCo Insurance, UK
-Working on Actuarial Pricing models and Data Science Projects in Insurance
-Power BI Dashboards for Claims analytics and Reserving
-LLMs Use Langchain and Open AI to read through PDFs of e-books on finance and insurance and ask queries to produce answers relevant to the analysis.
Lead Data Scientist, Tecxar Pvt. Ltd., Banking, India
-Working on ML models and dashboards to increase loan recoveries for loans.
-Brainstorm and develop ML use cases for improving recoveries for loan accounts.
-LLMs Using generative AI to produce info. Based on text search from data using ChatGPT with packages like OpenAI and Hugging face.
Data Architect - ML/Technical Lead
Koantek AI Cloud Pvt. LtdProject- Reverse Logistics, Market Mix Models, Anomaly Detection, Stock Price Prediction Model
Working in a tech. lead role managing a team of data scientists and engineers, responsible for designing the data architecture, pricing, data management, model development on cloud using AWS/ Azure/ Databricks. Working on AI/ML models for US based clients mainly from BFSI and Retail domain. Developing model using Python/SQL as a language in Databricks cloud. Develop business presentations and BI visualizations (Power BI/ Tableau) to showcase results and business outcomes
Responsibilities-
Retail: Reverse Logistics for clothing (UK) Design architecture for model deployment in AWS and predict the returns for particular brand/ region/ Colour/ Material/ Cost etc. for the clothing.
Pharma/life Sciences: Marketing Analytics (Korea) - Built an ARIMA model: Predict the future sales for different products. Automated and processed nearly 130 ARIMA models using a common python code & calculate the monthly carry over effect.
Finance: Stock Price Prediction Model (US) Design architecture, setup ETL and predict the stock prices for next 60 days and test the prices through different algorithms :Long- and short-term memory (LSTM), ARIMA/ Auto-ARIMA and FBprophet.
Successfully managed 3 data engineering work projects :E2 Migration PoC Migrate the data for the client from one workspace to the databricks workspace. Build an ELT process and Data pipeline Develop the same on databricks workspace.
Manager - Data Science and Analytics
PWC India- Assess & benchmark the firm for ML/AI to analyse for potential development opportunities.
- PoC - NLP Model to shortlist resumes for HR screening to save time and finances.
- PM Schemes (India) Analyse & develop an integrated Tableau dashboards for nearly 100 rural development Schemes (Team size: 6).
- Taxation Fraud Analytics (India) To identify tax evasion for medium and large-scale traders for the goods movement:
- Process and analyse the data using SQL to understand the reasons for cancellation, rejection, and unknown trader details.
- Perform clustering/ segmentation and Market basket analysis to analyse the pattern for movement of goods.
- Developed a Tableau/Power BI dashboard for ministries to analyse the observed anomalies and identify the traders.
Project- AI Strategy Plan (Dubai), NLP Models, Fraud Taxation analytics, Computer Vision, Clustering
Working in a Manager cum technical consultant role managing a team of 5-6 resources on client sites. plan and develop ML models and dashboards for taxation, PM rural development schemes. Fraud analytics using clustering to identify the loopholes for retail merchants in taxation develop an AI Strategy Plan for a client in Dubai for 6 months as a team of 4. This project involved benchmarking and assessing the organization, identifying the bottlenecks and recommending the AI Plan for the next 5 years along with developing 5 PoCs to showcase the abilities from AI. developed a couple of PoCs NLP Model for HR Analytics and computer vision to analyse and understand the traffic movement on expected bottlenecks in Dubai.
Responsibilities-
Actuarial Consultant
AXA Insurance- GLM Pricing models (UK & Turkey) -Develop Risk Premium Models for Motor & Property using EMBLEM, RADAR & Python.
- Risk Premium for Retail Motor Model the frequency, severity and propensity for different perils to develop a risk premium model.
- Model Validation and Backtesting Monitor the model and simulate the data to improve the performance of the model in the market.
- Gradient descent boosting method Use GBM to optimize the residual for the model to achieve a better accuracy.
- Developed Pricing models for Small & medium enterprises (SME) property for Turkey.
- Pricing Impact analysis Simulation of the model data to validate the model results in RADAR
Responsibilities-
Senior Consultant-Data Analytics
Allianz Insurance- Retention Modelling (Pet, UK) Logistic model to predict the customer retention.
- Develop customer retention model for cats and dogs for 2 products Yearly and life time cover for pets.
- Modelled for different pets and performed model validation using the hold out sample and other validation parameters.
- Commercial Branch & Broker Analysis (UK) Identify brokers for sales opportunities.
- Compare the brokers and different branches for Motor and Property across UK to analyse the growth opportunities. Output was the analysis along with a list of potential brokers and branches to be target for sales growth.
- Motor Loss Ratio Recovery (UK) - The project was carried out in 2 phases.
- Accident & U/W year based LRs, claim frequencies and severity by perils using the ladder method to develop monthly analysis.
- Analyse claims development pattern by different policy rating factors.
Responsibilities-
Data Analytics Assistant
RSA Insurance- Construction Pricing & Machinery Breakdown tool (UK, Technical Excellence Award) Developed market pricing tools with MI Dashboards (Team size: 3):
- Develop an Excel-VBA based pricing tool to calculate the benchmark, technical and actual premiums and help underwriters analyse the multiple technical adjustments for the large risks.
- Develop a summary tool for the management to analyse risks insured across geographies to prevent accumulation of risks.
- Quotes Pricing Analysis (UK) Developed an Excel-VBA based quotes analysis KPI dashboard.
- Developed a dashboard to analyse the quotes conversation based on value, volume, line of business and geography. Customer insights was provided in a dashboard PPT format automated to display across portfolio
- Portfolio Dashboard (HK), Industry Analysis (Singapore) & Migration KPI (Denmark)
Responsibilities-
Testimonial
Permanent and Contractual work roles
Client Feedback and testimonialsYou can find feedback on my LinkedIn profile under work samples as a PDF for contractual engagements and testimonials from the global leaders for whom I have worked in the industry.
https://www.linkedin.com/in/mohit-bansal-7227bb34/
Major Projects
LLM - ChatGPT genrative AI - Personalized experience
Tecxar Pvt. Ltd.- Business Goal - To customize the ChatGPT for internal organizational purposes: Create an internal chatbot for helping employees understand the company policies and ask questions, provide information on products to customers through online customer care chatbot.
- LLMs - Using generative AI to produce info. based on text search from data using ChatGPT with packages like OpenAI and Hugging face.
- Inputs were the company policies and product brochures and schema to answer the queries from the users.
- Use LLMs for market analysis, benchmarking studies, backend for chatbots, answer questions from students for educational benefits.
Deep Learning CNN Model - Banking
Tecxar Pvt. Ltd.- Business Goal: To develop a data science model to predict the probability of loan recoveries for the loan defaulters.
- Product: Micro-finance loans, Vehicle (2/3/4 wheeler) and Secured loans (Gold)
- ML Algorithms: We used a Logistic model with GBM to improve accuracy initially but data processing became a challenge as data got heavy as accounts increased.
- Later, we used a deep learning CNN model to train the data and predict probability of success done in 2 phases to achieve best results.
PoC: LLM - Generative AI - Answers from ebooks
Tecxar Pvt. Ltd.- PoC: Use Langchain and OpenAI to read through PDFs of e-books on finance and insurance and ask queries to produce answers relevant to the analysis.
- Purpose: PoC was done for education sector as it is in demand with Chatgpt's introduction.
- Data input: 100 ebooks used as data input.
- Output: Both text and Audio ability based on the questions asked by the user.
- Similar to a chatbot but with limiting the information to provide answers.
Reverse Logistics, Market Mix Models, Anomaly Detection, Stock Price Prediction Model
Koantek Cloud and AI Services Pvt. Ltd.- Working in a tech. lead role managing a team of data scientists and engineers, responsible for designing the data architecture, pricing, data management, model development on cloud using AWS/ Azure/ Databricks.
- Working on AI/ML models for US based clients mainly from BFSI and Retail domain. Developing models using Python/SQL as a language in Databricks cloud.
- Develop business presentations and BI visualizations (Power BI/ Tableau) to showcase results and business outcomes.
Recommendation system using NLP - Start ups vs Investors
Tecxar Pvt. Ltd.- Business goal: To match and generate top 5/10 recommendations for the start ups with investors and vice versa.
- Solution suggested: An NLP model with content based recommendation (RC) system.
- Data input - Data for 1 year for companies and investors.
- Techniques: NLP pre-processing + Rule based algorithm + cosine similarity algorithm for recommendations.
AI Strategy Plan (Dubai), NLP Models, Fraud Taxation analytics, Computer Vision, Clustering
PriceWaterhouse Coopers (PwC) Consulting India- Working in a Manager cum Technical consultant role managing a team of 5-6 resources on client sites.
- Plan and develop ML models and dashboards for taxation, PM rural development schemes.
- Fraud analytics using clustering to identify the loopholes for retail merchants in taxation.
- Develop an AI Strategy Plan for a client in Dubai for 6 months as a team of 4. This project involved benchmarking and assessing the organization, identifying the bottlenecks and recommending the AI Plan for the next 5 years along with developing 5 PoCs to showcase the abilities from AI.
- Developed a couple of PoCs NLP Model for HR Analytics and computer vision to analyse and understand the traffic movement on expected bottlenecks in Dubai.
ML Models in Insurance- GLM Pricing Models, Model Revision and Pricing
AXA Insurance Group- GLM Pricing models (UK & Turkey) Develop Risk Premium Models for Motor & Property using EMBLEM, RADAR & Python.
- Risk Premium for Retail Motor Model the frequency, severity and propensity for different perils to develop a risk premium model.
- Model Validation and Backtesting Monitor the model and simulate the data to improve the performance of the model in the market.
- Gradient descent boosting method Use GBM to optimize the residual for the model to achieve a better accuracy.
- Developed Pricing models for Small & medium enterprises (SME) property for Turkey.
- Pricing Impact analysis Simulation of the model data to validate the model results in RADAR.
Education
M.Sc. Applied Operational Research
South Campus, University of Delhi (2012)B.Sc. Mathematics (Hns)
Ramjas College, University of Delhi (2010)CBSE (AISSE & AISSCE)
St. George s School
Certifications
- Credential ID : In Progress
IBM Full Stack Software Developer Course
(Jan, 2024)