Anchit Pandey

Graduate Student Developer

Center for Systems & Software Engineering•  May 2020 - August 2020

• Implemented Code2Seq research paper in Python TensorFlow 2 to construct baseline model with accuracy of 82% • Built transformer model in Python & merged it with Code2seq baseline improving accuracy to 85% • Devised LSTM & GRU units in Python and merged it with baseline model yielding accuracy of 81%

BI Developer Intern

PricewaterhouseCoopers (PwC)•  January 2019 - July 2019

• Coded customized ETL framework in C#.NET reducing data extraction runtime from CRM to SQL-Server by 30% • Created 3 data models in SQL-Server & 3 ERD in Lucidchart for client project & POC. Refined business logic of model • Coded 11 .NET scripts to perform data validation on attributes of relational table for client project • Prepared 4 dashboards using Power BI for client & POC with slice & dice, drill down & drill through functionalities

Teaching Assistant - Java Programming Lab

Birla Institute of Technology & Science Pilani, Dubai Campus•  September 2018 - December 2018

• Guided students to write efficient code for algorithmic problems • Coded 70 Java reference solutions for greedy, graphs, Dynamic Programming, divide & conquer problems utilizing data structures like Hashmap, BST, stack, queue, suffix-tries, Heap, Disjoint Set • Debugged errors & optimized students code by enhancing both runtime and memory consumption

Deep Learning Developer Intern

Esri•  October 2017 - December 2017

• Executed SIFT, SURF & ORB algorithms in Python OpenCV to detect terrain features in ARCGIS maps with 87% accuracy • Wrote SQL procedures to query 10000 images in MySQL & return 10 closest matched tiles for input image • Built & trained CNN in Python for identifying airport by detecting airplane cluster in ARCGIS images with 81% accuracy • Preprocessed 10000 image using Python script by splitting image to grid tiles, applying padding masks & Gaussian kernels

Software Development Intern

Intellect design arena•  June 2016 - August 2016

• Developed Vendor Portal web app in Angular and constructed REST APIs in Express.js for authentication, committing transactions & displaying historic data • Devised 13 Angular components, services, routers, models & Auth-Guards for each logical unit of project • Consumed REST APIs in Angular app & integrated internal software of organization with web app