Manish Nanwani

+91-8443695332 | nanwani1@gmail.com |

SUMMARY
A curiosity-driven data scientist, eager to leverage machine learning and data analytics to extract meaningful insights,
make informed decisions and solve challenging business problems. I ensure to contribute with my knowledge, logical
thinking and analytical skills towards the consistent growth and development of the organization, and enhance my
experience through continuous learning and teamwork.

WORK EXPERIENCE
Pivotchain Solution Technologies
Data Scientist June 2018-Present
• Worked on a PoC for a leading Dubai Telecom Company, involving Descriptive and Exploratory Analysis of the
Trouble Tickets incidents raised for a period of 6 months, identified patterns, gathered meaningful insights into key
issues, and created dashboards for visualizations.
• Automated Speech Recognition, which converts speech to text, for various domains. Data Collection by scraping
through various financial websites and international news websites, to be fed to the language model.
• Conversational Language Model. Dataset Preparation, which included extracting and fragmenting audio (raw data
samples) from movies and TV Series and mapping them to their respective subtitles (target/labels).
• Web Crawling and Automation of processes for a Dubai-based Medical Healthcare Service Provider, with the objective
to authenticate the validity of Insurance Claims by consolidating results from various Insurance Company Portals.
• Extracting information of data fields from various Government documents, using Optical Character Recognition.
• Face Detection and Identification with the Person of Interest, both in Video streams as well as from large collection of
target images, using Computer Vision techniques and Deep Learning.
Alfa Laval India Pvt. Ltd.
Graduate Engineer Trainee 2016 - 2017
• Handled large databases of Marine Products installed on both the Indian Navy vessels and Commercial Fleet globally.
• Collected data for the Customer Installed Base pan-India, which was later used for grouping the customers region-wise.
Tata Consultancy Services (TCS)
Summer Internship May - June 2015
• Worked on a Business Intelligence Analysis Project on Tableau - Creating a dashboard for a top engine manufacturing
company, for graphically visualizing their financial targets, as well as other parameters.

EDUCATION
Aegis School of Business, Data Science, Cyber Security & Telecommunications 2017-2018
Post Graduate Program in Data Science, Business Analytics & Big Data in association with IBM
College of Engineering, Pune (CoEP) 2012-2016
Bachelors of Technology, Mechanical Engineering - 8.5 CGPA
Nowrosjee Wadia College, Pune 2010-2012
Higher Secondary Certificate (HSC) - 91%
The Bishop’s School, Pune 1999-2010
Indian Certificate of Secondary Education (ICSE) - 89%

SKILLS
• Statistics – Various hypothesis testing, estimation, probability theory, time-series analysis, statistical modeling.
• Machine Learning – Algorithms for Regression (Linear, Logistic), Classification (Decision Trees, Random forest,
XGBoost,SVM, Naïve Bayes, k-NN),PCA, Clustering(k-means, Hierarchical).
• R - Implemented ML Algorithms, used packages like dplyr,glmnet, ggplot, caret,boruta, missForest, mice, dummies
• Python - ML Algorithms, using packages like Numpy, Pandas, SciPy, Scikit-Learn, Statsmodels, Matplotlib, Seaborn
• Natural Language Processing – Text Processing, Web Scrapping, Sentiment Analysis, Regular Expressions, NLTK
• Deep Learning – Neural Networks, RNNs, CNNs, LSTMs; tensorflow, keras.
• SQL - Performing Basic Queries, Sub-queries, Joins, Aggregation, Statistical Functions.
• QlikSense and Tableau – Data Visualization, Business Intelligence, Forecasts, Tables, Charts, Dashboards
• Big Data – Basic knowledge of Hadoop, Map Reduce and Spark, along with all the other tools of the eco-system.
PROJECTS
• Meetup Group Analysis (Project Link) Tools: Python: selenium,nltk, pandas
- Worked as a freelancer on an independent project for a San Francisco based Comic Company, for a duration of 4
months. It involved Exploratory and Descriptive Analysis of various Meetup-groups in the comics domain globally from
the website www.meetup.com and deriving meaningful insights from it.
- Ideated and executed Web-Scraping the meetup website to further collect and gather data for every group, its founders,
members, location, interests of the member of the groups, genres of groups, other groups a particular member is part of.

• Pover-T Tests: Predicting Poverty of a Household in Different Countries Evaluation: Log-loss: 0.215
Rank: 279 out of 2000+ participants (Project Link) Tools: Python
Predict whether a given household in a given country is poor or not. Survey Data was provided by The World Bank
Development Data Group from 3 countries, both at household as well as individual level, hosted by Driven Data. Total
12 CSV files present: 300 columns and 12K rows at household level and 43 columns and 60K rows at individual level.

• Missing Person Detection in a CCTV footage (Project Link)

Algorithm: Convolutional Neural Network Tools: Python, tensorflow, keras, opencv, dlib, flask
Building a solution which checks whether a particular person in the input photo is identified in a given set of CCTV
footage. This automates the manual search process, and returns the frames and timestamp of that person, if present. The
other use case is identification of a wanted person in the CCTV footage. Deployed the model on Flask

• AirBnb -New User Destination Country Prediction:(Project Link) Evaluation: Accuracy:0.672

Applied Multinomial Classification Algorithms, combining data from 4 CSV files, having user details, demographics,
signup properties, 10 lac rows of web session records for the users. Main dataset had 2.2 lac rows train, 62K test data.

• West Nile Virus Prediction: (Project Link) Evaluation: AUC: 0.656

Predicting the presence of West Nile virus in the mosquitoes across the city of Chicago, in Python. The dataset was
provided by Chicago Department of Public Health. The main dataset had 11K train rows, 1.2 lac rows test data ,and 13
features, like mosquito species and number, geo-location address data, using data from 2 other files containing GIS data
on the Spray Area and the weather conditions within Chicago.

• Mobile Phone SMS Spam Filtering: (Project Link) Evaluation: Accuracy: 0.97
Developed a classification algorithm, to filter mobile SMS spam, using the Naive-Bayes Algorithm. The dataset
consisted of a corpus of 6K labeled SMS’s; applied text analytics and language processing techniques to train the model.

• Employee Churn/Attrition: (Project Link) Tools:R:ggplot, RMarkdown

Data Exploration and Visualization of employee attrition trends in an organization from British Columbia, Canada. The
dataset of 50K rows, was a 10 year employee history, consisting of titles, departments, store locations, years, city, etc.

• Northwinds Traders Sales Visualization: Tools: Tableau, QlikSense

Created dashboards on 3K rows dataset containing the sales, suppliers, customers, employees, products, orders, shippers,
and countries. The dashboard gives profit/loss indications, sales distributions, delivery status, and other insightful KPIs.

P.S.: Have worked on numerous data science hackathons and machine learning projects, uploaded on github:
https://www.github.com/manishnanwani