Hey! I am

Shivam Sharma

I'm a

About

About Me

  • Data Scientist Professional offering 1.5+ year of industrial exposure in Machine Learning, Deep learning, Data Analysis and Computer vision
  • Also, worked in data engineering, data analysis, ETL process creation using PostgreSQL, Python and R
  • Name: Shivam Sharma
  • Date of birth: January 23, 1995
  • Address: E-1305, Arihant Arden, Greater Noida, UP, India
  • Zip code: 201306
  • Email: shivam17147@iiitd.ac.in
  • Phone: +91-9910902830
  • Orcid ID

0 Project complete

Download Resume

Education

2017-2019

M.Tech in Computational Biology

Indraprastha Institute of Information Technology Delhi

Indraprastha Institute of Information Technology, Delhi is a state university located in Delhi, India. It is research-oriented with a focus on Computer Science and allied areas.

2012-2016

B.Tech in Information Technology

Cluster Innovation Centre, Delhi University

Its an institution of eminence exposer to technology, new studies and renowed faculty with the moto "Dissolving Boundaries and Evolving Senses"

Experience

Jan 2020 - May 2020

Data Scientist

Knowdis Data Science LLP
  • Performed real time objects recognition using Mask R-CNN
  • Worked on YOLOv3, faster R-CNN, Resnet101
  • Deployed the deep learning models in a flask application
  • Optimised the performance upto 80% for real time prediction
March 2019- Dec 2019

Data Scientist

Circle Of Life Healthcare Pvt Ltd

Deliverable:

  • Data analysis and ML model implementation in R and Python
  • Data engineering, providing assistance in product deployment and building new feature in the product.
  • Antibiotic Panel Prediction using recommendation system
  • Outlier Detection for Bacteria Prevalence
  • Trend analysis for detection of increase in antibiotic bug resistance
  • APIs and ETL implementation in Python with PostgreSQL
  • Provided support in onsite deployment of the product Zevac (inspired from my M.Tech thesis)
June 2016 - June 2017

Project Fellow

Indian Statistical Institute Delhi

Implemented efficient public-key cryptosystem based on an elliptic curve from scratch, relying only on large-number arithmetic libraries like GMP NTL & MPI in C/C++.

Skills

Python

100%
100%
Last week
100%
Last month

Machine Learning

90%
28%
Last week
60%
Last month

Deep learning

90%
100%
Last week
100%
Last month

Flask Web Development

70%

MATLAB

70%

R

95%

Django

90%

Tensorflow

90%

Keras

80%

Reseach Papers

2019

A Game Theoretic Model of Deceptive Ambush as Countermeasure for Habitat Selection in Cross-Border Infiltration (Springer)

Indraprastha Institute of Information Technology Delhi and DRDO Delhi

Cross-border infiltration presents a national concern. Defense and monitoring of border areas require the inclusion of a wide range of tactics to interdict infiltrators. In an infiltration situation, human habitats along the infiltration route present numerous advantages to an infiltrator. The security personnel can adopt various strategies with a habitat to ambush the infiltrator since it is highly likely for the infiltrator to visit a habitat. The habitat acts as a proactive ambush site for the defender and therefore presents him with a high probability of deceiving the infiltrating agent. In this paper, a two-player zero-sum game is proposed between a defender and an infiltrator for habitat selection under the threat of ambush with and without deception. Both players apply various strategies to maximize their benefit out of the options available to them and adjust to each other’s options over the play only to eventually settle on an optimum strategy. The two players determine the tactical options available to them. The ambusher assigns a numerical value (Measure of Effectiveness, MOE) to each possible outcome by judging the potential gain or loss from an encounter. With the proposed game models the players can check to see if the outcome of the encounter in either deceptive or non-deceptive scenarios favors him or the opponent and what optimum strategy is possible with a desirable outcome. To the best of our knowledge, the present paper is the first application of game theory to human-human interaction similar to a predator-prey scenario for habitat selection.

2015

Extension of Josephus problem with varying Elimination Steps (DU Journal of Undergraduate Research and Innovation)

Cluster Innovation Centre, DU

This paper is an effort of expounding the recursive formula for the Josephus problem (which is the essence of the popular game “Akkad Bakad Bambai Bo” played in Indian households). The academic extension of this game has been to deal with the cases where elimination steps vary with each turn. Josephus problem is essentially heralded with objective of finding the position of the player who survives the game. Most of the available literature deals with the problem of finding the survival’s position when sequential elimination of persons takes place in steps for circled people. There is a simple algorithm known to solve this problem; however the rigorous procedure can be replaced with a formula. The nature of this formula can be either recursive or non-recursive; both of which comprise this paper. We have extended the problem to the situation wherein the elimination step varies with iterations.

I'm Available for freelancing

Professional Data Scientist with 1.5+ years of experiece in Data Analysis, Machine Learning and Computer Vision

Hire me

Contact

Contact Me

To Know about my research and more about my drop a message to me. I will be glad to connect with you.

Address

E-1305, Arihant Arden, Greater Noida, UP, India

Contact Number

+91 9910902830

Email Address

shivam17147@iiitd.ac.in

Linkedin

Profile