MINAL PATIL

Greensboro, North Carolina +1-336-***-**** adujcs@r.postjobfree.com

https://www.linkedin.com/in/minalrajesh

MS (Data Science) grad. student available for full time employment starting Feb ‘23 PROFESSIONAL SUMMARY

IT and HR professional with around 5 years of experience in software programming and human resources domain. Worked as Java programmer for IT based companies in India and on ERP -SAP as a functional consultant for a power generation company based in Mumbai, India. Currently pursuing master’s in data science to make a fresh start in the exciting field of Artificial Intelligence. EDUCATION

• Pursuing master’s in computational data science & Engg. at North Carolina A&T State University, USA. Expected graduation: May ‘23. 1st year GPA – 3.94

• Master’s in management studies (HR), Mumbai Univ, India. (70.1%) 2008 – 2010. GPA– 3.90

• Bachelors in Info. Technology, Pune Univ., India. (64%) 1999–2003. GPA–3.13 TECHNICAL SKILLS

• Programming Languages- Python, R, Java

• Databases: SQL workbench

• Data Visualization

• Machine learning libraries: sklearn, pandas, numpy, matplotlib, keras.

• Deep Learning - Natural Language Processing (NLP) basics, Image Processing (CNN)

• Understanding of RNN, Transformer, BERT.

• Ubuntu- Windows sub-system for Linux, BASH

CERTIFICATIONS

• Python Basic- Hacker Rank certification. Jan 2023

• PCAP – Certified Associate in Python Programming – by Python Institute May 2021

• PCEP – Certified Entry-Level Python Programmer- by Python Institute. April 2021

• Human Resource Business Professional (HRBP) certified by HRCI. Sept 2014

• SAP HCM Evaluation Certification- From L&T Infotech, India. Jan 2011 ACADEMIC CASE STUDIES

1. Simple movie recommendation system with truncated SVD algorithm Objective: to predict movie ratings given data as user, movie, rating table. Analysis: Any kind of recommendation problem involves data in millions of records, but the data could be sparse. Processing huge sparse data matrices need large compute & cost. Model: truncated SVD is used to decompose this matrix which could have been of size of 943 x 1681

= 1,585,183 to a reduced size (943, 10), (10,) & (1681, 10) and predicted the ratings using Collaborative Filtering with a MSE of 0.8479. This work can be used for any kind of recommendation system like user, item on amazon website. As an extension to existing work, we might consider using additional features apart from the user is and the item id.

2. Applying LSTM on Donor Choose dataset:

Objective : Predict if a project request received on DonorsChoose.org will be approved/ rejected Analysis: DonorsChoose.org receives hundreds of thousands of project proposals, from across the schools in USA, which can’t be screened manually to decide whether they should be processed further for approval. The process must be automated. Models applied: The text, categorical and the numerical data is preprocessed accordingly. The text is converted into numerical data with the help of GLOVE model and the embedding layer to be fed into the LSTM model. Similar embedding layers formed for all the categorical and numerical layers.

Other models like Decision Tree, Naïve Bayes are also tried. LSTM’s provides the best results with a test AUC of 0.759 compared to Decision Trees (0.612) and NB(~0.646).

The application of this model is general purpose for many binary classification tasks. The accuracy may still be improved with the state-of-the-art techniques like transformers. 3. Other implementations:

• Custom implementation of TFIDF vectrozer: Implemented TFIDF vectorizer from scratch. The results are comparable to the sklearn implementation of the same.

• Applied bootstrap sampling method in Random Forest to introduce randomization for regularization.

WORK EXPERIENCE

SAP HCM Functional Consultant (LTI -Larsen & Toubro Infotech – India) (Jan ‘11 - Aug ‘12)

-Involved in SAP implementation of HCM module in MahaGenco (Power Generation Company, Maharashtra State-India, -Was one of the team members for mapping client's requirement into the SAP system.

-Provided end user training to the employees.

-Successfully completed data migration to new system. HR consultant (Phi Consulting- Jun ‘10 - Dec ‘10)

Worked on various HR related activities including recruitments, in-house training, compensation and benefits.

Programmer-Java (Base Information Management Pvt Ltd – Apr’05 - Dec ‘06)

-Involved in supporting WEBITM with SQL, generation of front-end screens.

-Providing end user training and support.

-Implemented 'Project Tracking' module in Sun Pharmaceuticals. Trainee Software Developer (Vanquish Info net pvt ltd. - Oct ‘03 - Jan ‘05) Worked as trainee software developer on various web portals like College Information System in microsoft technologies like ASP3.0, SQL Server 2000, JavaScript, etc.

Contact this candidate