Patents Pending:Rohan Reddy Mekala, Matthias Muehle, Adam Porter, Mikael Lindvall and Michael Becker: “Artificial intelligence based algorithms for prediction of diamond crystal growth” U.S. Provisional Pat. App. No. 63029177, filed 22nd May, 2020.Rohan Reddy Mekala, Adam Porter and Mikael Lindvall: “Metamorphic filtering of black-box adversarial attacks on multi-network face recognition models” U.S. Provisional Pat. App. No. 63002728, 31st March, 2020.Publications:Rohan Reddy Mekala, Gudjon Einar Magnusson, Adam Porter, Mikael Lindvall and Madeline Diep: “Metamorphic Detection of Adversarial Examples in Deep Learning Models With Affine Transformations” 4th IEEE/ACM International Workshop on Metamorphic Testing (MET) at ICSE 2019. Rohan Reddy Mekala, Adam Porter and Mikael Lindvall: “Metamorphic filtering of black-box adversarial attacks on multi-network face recognition models” 5th IEEE/ACM International Workshop on Metamorphic Testing (MET) at ICSE 2020. Rohan Reddy Mekala, Asif Irfan, Mikael Lindvall, Adam Porter and Eduard C. Groen: “Classifying User Requirements from Online Feedback in Small Dataset Environments using Deep Learning” Pending approval

BIG DATA WITH SPARK (PySpark + SparkSQL + Pandas/Python) [Pocket Security]♣ Developed framework to redo ETL for real time data processing (> 2 million records) using PySpark on in-house Spark cluster set (Extraction from Google Datastore)♣ Improved LOAD time by 30% by developing algorithms to overcome repetitive re-creation of HDFS datastore (while preserving table schemas) due to the immutable nature of RDD structures♣ Improved report creation speed incrementally by distributing the extract phase for individual parquet tables across clusters using Pandas Dataframes and Spark NLP (NLTK/Stanford NER/Spacy - Python) for STATE MACHINE REQUIREMENT DOCUMENTS [John Deere]♣ Developed application to kick start a reversal of the manual aspect of reading voluminous unstructured requirement documents used to create state machine diagrams ♣ Coded extract custom entities using semantic analysis with chunking, part of speech tagging, context free grammar regex and relation algorithms ♣ Created custom models for the three main NLP pipelines using RNNs with LSTMs to achieve optimal grammar forms and their dependency trees♣ Achieved 87% accuracy developing algorithm to extract relations between target entity types (for an ontology framework)DATA VISUALIZATION AND MACHINE LEARNING (R, Shiny R) [John Deere]♣ Feature engineering to format semi structured data to over 90 independent variables ♣ Coded classification imbalance improvement and machine learning models with tuning algorithms to zero in on a framework improving accuracy over the previously used version by over 10% (using Gradient Boosted ensembles)♣ Built comprehensive web based user dashboards for data modelling/visualization and predicted code defects using dynamic manual user inputs across multiple target sources AMAZON LEX + AWS LAMBDA + AWS API Gateway [Dosesmart]♣ Created application to access data transmitted by the IOT device enabling

IMAGE BASED FEATURE DETECTION (Python,OpenCV, Numpy) FOR WIRE HARNESS MICRO-FEATURES [SAFRAN Aerospace]♣Developed algorithm to identify dynamic pattern detects in aviation grade wire harnesses, employing Gaussian Mixture Models, KNN, image quantization, edge/shape detection, information loss approximation, feature detection and CNNs. ♣Achieved classification accuracy of 99%, improving the previously presented version by 64%

Founded a tech analytics firm based on web/mobile app development, search engine optimization, social media analytics and comprehensive digital marketing solutions♣ Worked till date with over 300 clients and cross sold web/mobile app and digital marketing projects worth $200K to existing and new clients♣ Created affiliate marketing funnels as advertiser for websites/landing pages of large scale product models like Fiverr, Digitalocean, Amazon and others, generating over 500,000 targeted clicks to corresponding products♣ Automated SMTP and bulk email management solution, scale tested to handle 15k+ campaigns in parallel♣ Started a technology magazine - OnlineSearch.net, to promote viral social media campaigns around creative content in the entrepreneur/technology niche

Founded and led Guidetrip.com to become one of the largest providers of curated tours in the world.♣ Selected as one of the 11 best global start-ups, by Global Super Angels Forum (GSF), one of Asia’s biggest technology accelerators; Raised seed funding of US$ 50000 from investors in India, Mauritius and Ukraine♣ Worked for 6 months in USA and Canada, to promote GuideTrip in the startup community across major cities♣ Built the entire initial technology for the website using Yii Framework, Javascript, Jquery, Angular.js and SQL from scratch♣ Managed 20+ employees and 200+ interns in spearheading SEO/SEM/Social Media automation initiatives, driving Guidetrip to over 20,000 travel experiences and 9,000 tour guides

Atlogys is a CTO (Chief-Technology-Officer) level IT Consulting firm that provides 'turnkey' software development services♣ Co-ordinated over 20 people in developing/managing more than 10 large scale technology projects based on website/mobile app development♣ Planned and developed a recruiter model website for SL8Z.com, which went on to raise VC funding in a matter of months♣ Managed teams at Guru.com and Atlogys to build project plans for Guru.com and single handedly developed a front-end architecture for the website, targeting reduction in bounce rate by 40% ♣ Coded a frontend JavaScript framework called brucelist.js for Guru.com to handle dynamic data binding and templating ♣ Worked towards online object modelling using the 3D graphics framework d3.js

ISSA (Institute of Systems Studies & Analyses) is the nodal System Analysis Laboratory of DRDO specializing in System Analysis, Modelling and Simulation of Defence Systems as well as development of Computer War games, Performance Evaluation and Systems Reliability Studies.Analysed and solved a real life problem in problems tracking fast moving objects like satellites, shuttles etc when using radars and sensors. Proposed, analysed and proved the efficiency of Kalman filters as the best solution by using code to simulate object motion and estimate real time tracking performance on a 2D level.