Sayan Mukhopadhyay

Sayan Mukhopadhyay Kolkata, India - •-• LinkedIn • Facebook Manager | Architect | Consultant | Analyst | Scientist | Author Multifaceted technical career with 13-year track record of innovation and success in corporate settings. Technically sophisticated professional, with extensive expertise in the fields of machine learning, cloud computing and data-driven product or solution development extending the horizon towards IOT and cyber security. Achieved proven success in the development of optimal systems that deliver tangible performance improvement. Skilled trainer and project leader; able to direct multiple tasks effectively and master innovative software and tools. A selection of key projects and accomplishments includes Managed and developed highly effective analytical solutions for a system receiving 100 million new records on a daily basis on behalf of a leading online advertising company. Led the development of an enterprise network management system, involving complex bug fixing and development of new features in one core of heterogeneous, distributed codes. Designed and developed a parser for FIX format file in C++ that improved efficiency ten-fold in comparison with Unix grep command; to support high frequency trading servers of a major investment bank. Strong research background, including three international publications, one applied patent and completion of MS by Research in Computational & Data Science and BEng in Electronics & Instrumentation. Core Competencies: Data Science / Artificial Intelligence High Performance Computing Full Stack Product Development Problem Resolution / Troubleshooting Client Relationship Management Solution / Technical Architecture Business Intelligence & Big Data Agile / Rapid Application Development / ITIL / ITSM Algorithmic Trading / Intelligent Infrastructure Motivational Leadership Professional Experience Sulvo, Remote working Head - Machine Learning & Infrastructure, 11/2016 – Present Tools: AWS, Node Js, python, tensorflow, Google cloud platform, falcon, redis, mongo, sqlite, RNN based predictor Building from scratch a real time deep learning based patent pending algorithmic strategy manager to maximize site revenue for online advertisement platform Sulvo selected for Google start up mentoring program. Ganit Mistry, Kolkata CEO, 01/2017 – Present Working for one of the oldest and largest algorithmic trading organization in India. Win a sales target prediction projects from one of the largest FMCG Company in India through vendors. Working for loyalty reward strategy engine for one of the largest business house of India. Abzooba, Kolkata, India Manager – Data Science, 6/2015 – 11/2016 Tools: Java, Play, Mongo, Python, Neo4j, Angularjs, Stanford NLP, Beautiful Soap, tensorflow, mem2n, AWS Promoted to the role of Data Science Manager and Project Lead for key client accounts. Lead a team of seven staff members to deliver sophisticated projects within budget and ahead of schedule. Direct the entire project lifecycle from the initial design and planning phases, through to progress / risk monitoring and implementation. Core member of the leadership team, providing strategic input to assist CXO's in critical decision making. Key Projects: Successfully designed analytic solutions (GUI to ML model) within a specified client framework in addition to establishing a credit score model using social media data. Designed and implemented optimal systems, such as dialog, security, auto-complete and feedback modules for retail e-commerce operations and a Topical Crawler for web harvesting. Developed and deployed a Health Insurance Claim Status prediction System via HIPAA 835 & 837 data as well as vendor selection and price optimisation systems utilising transactional data. TCG Digital, Kolkata, India Tools: Python, Java, Mahout, Hadoop, Hive, Elastic Search, Kibana, Sentiment Analysis, Neural Network Predictor Data Scientist / Sr. Consultant - Analytics & Big data, 7/2014 – 6/2015 Critical role, applying data science methods, processes and systems to troubleshoot analytics and complex bigdata issues for diverse clients. Built positive client relationships, with direct involvement in pre-sales activities. Key Projects: Played a lead role in a sentiment analysis project focused on brand development as well as a Neural Network based Predictor for passenger loads for two clients in the aviation sector. Delivered a solution to address data errors for a leading electronics manufacturer / distributer, via the parallel implementation of KMean clustering using Needleman Wounch sequence matching score. Mphasis (HP Group), Pune, India Technical Architect – Transaction Monitoring System, 10/2013 – 7/2014 Tools: Nagios, Cacti, Ninja, Merlin, Lucierna, Combodo-itop, Jasper-Report, Sumologic, Event correlation, Topology Auto-discovery Fully accountable for the build, development and deployment of a complex transaction monitoring system based on available open source stack. Provided optimal service to nine key clients, including Coca Cola, Verion and RPG. Performed requirements gathering and feasibility analysis, then developed robust project plans and roadmaps. Managed, motivated and coached the technical team, with 12 members, to achieve all objectives. Key Projects: Designed and implemented enhanced system architecture, including a framework for Nagios Active Checks, recursive approval method for change management in ITSM Integration of netflow and Nagios core via netflow collector and Nagios check plug-in. Swiftly identified / fixed bugs, such as a data miss issue of cacti with 'polar table in not empty' exception. Wedoria (ABP Group), Kolkata, India Lead Architect - Data Mining, 5/2013 – 8/2013 Tools: Bass Model, Collaborative filtering, Mechanism design, MySql, PHP, Java, Python, Storm Lead Architect, tasked with creating a revenue model using a stochastic Bass model of user growth / influence. Formulated architecture to support a recommender system and design mechanism incorporating collaborative filtering and linear programming. Effectively managed two direct reports; Data Scientist and Data Engineer. Pubmatic, Pune, India Senior Technical Analyst - Machine Learning, 10/2013 – 7/2014 Tools: Java, Perl, Shell, Python, Hadoop, MapReduce, Hive, Storm, MySql Applied predictive approximation techniques for big data queries to enhance the Hadoop platform in addition to designing and deploying a new estimation and prediction framework for query handling. Conducted R&D, established a unit testing framework and applied Agile methodologies to ensure seamless project execution. CA Technology, Hyderabad, India Senior Software Engineer - Network Management System / Data Centre Operations, 9/2011 – 4/2012 Tools: C++, Java, Perl, SNMP, Barclay DB, Gmake, Tomcat, Rest-Ful Web Service, RTC Vital role as senior member of the Spectrum Backlog Reduction team, tasked with the development, maintenance and support of a large-scale, complex data centre operation domain. Performed bug fixing, code reviews and testing, then designed and deployed features to enhance performance, quality and security standards. Delivered 2nd Level support to clients and mentored junior members of the team. Key Projects: Successfully implemented numerous product features, such as an ignore character case in seek command enabling users to find a model handler with given condition. Swiftly resolved complex bugs, including a GUI fault due to race condition between front end thread for servlet request and backend thread of server response. FairFest Media Limited, Kolkata, India Technical Manager - Patent Analysis / Loyalty Reward Program, 11/2010 – 9/2011 Tools: ASP/C# .net, Nhibernate, WCF, C++, Python, Poco, NVP/SOAP/Rest, AJAX Conducted patent analysis, marketing and development activities within the poverty, environment and security areas. Acted as a focal point of contact on technical issues and mentored the 14 members of the project team. PayPal, Chennai, India Senior Consultant - Application Development, 3/2010 – 11/2010 Tools: Apache, Red-hat Linux, C++, Swig, Python, Java, ClearCase, Active MQ Strategically managed and developed core payment technology, including the optimisation of system architecture, framework and functionality. Oversaw code review, debugging, profiling and unit testing activities, with four direct reports. Promoted an environment of continual improvement and best practice. Key Projects: Established an AMQ framework that recorded corresponding error numbers in the Common Application Logger based on errors related to SQL. Developed a remove range method in Vector class for the PayPal developer community and applied swig generate python / java interface of admin-auth module to cache authentication data in a local server. Credit-Suisse, Pune, India Technical Consultant - Investment Banking, Risk Analysis, Data Centre Operations, 6/2008 – 2/2010 Tools: C++, C# .net, Perl, Shell, Python, Oracle, Sybase, Linux, Solaris, Tibco-EMS Multifaceted role, providing technical consultancy across risk analysis, algorithmic trading and infrastructure development as well as acting as Rapid Application Developer for high-frequency trading. Key Projects: Delivered an application support project involving the modification and enhancement of a trading time risk analysis tool which greatly improved risk management and control. Successfully designed and implemented a distributed, divide and conquer Real-time Log-file Parsing Framework for use in high-frequency trading activities. Prior experience as Software Engineer (1/2000-12/2003) in Total Computer System for different academic clients. Technical Proficiencies Programming: C++, Core Java (6.0), C# .net(3.5), Shell, Python, SQL, Node-Angular, Android, UML 2.0 Tools: Socket, IPC, RPC, ACE, Poco, HTTP, REST, SOAP, Play, Flask, WCF, Unix (GDB, Gprof, Makefile), HPC (MPI, Open MP, POSSIX, CUDA), Selenium, Active-MQ, Tibco-EMS, Zero-MQ, git, svn Analytics: Matlab, R, Octave, Oracle, Sybase, Mysql, Sqlite, Elastic Search, MongoDB, neo4j, Hadoop, Mapreduce, Hive, Mahoot, Storm, Spark, Gemfire, MS BI, Pentaho, D3, C3, Google Visualization API, Map API, Kibana, SAS Certified Base Programmer, SAS Systems Administration: Linux, Windows, Solaris, AWS, Google cloud, Virtual Box, Router, Switch, SevOne, Nikson, Corvil, Ganglia, Spectrum, E-Health, Nagios, Cacti, Lucierna, Asterisknow, socat Security NMAP, SearchSploit, SQLMap, BurpSuite, armitage, Kali Linux Education and Credentials MS by Research in Computational & Data Science, IISc, Bangalore, India BE in Electronics & Instrumentation, Jadavpur University, India, CSIR Fellow in Computational Solar Geometry Publications: Utility of variance of difference as a distance like measure in synchronous time series microarray data clustering, Sayan Mukhopadhyay, Dr. Debnath Pal, BMC Bioinformatics Fully parallelizable algorithm for molecular dynamics of protein’s soft docking simulation, Sayan Mukhopadhyay, Dr. Debnath Pal, Journal of Computational Biology Advance Data Analytics using Python, Apress(Springer), Sayan Mukhopadhyay (Book) Other Information: Selected for National Math Olympiad; Ranked 352 in West Bengal Joint Entrance Examination (Engineering); All India rank 50 in Graduate Aptitude Test in Instrumentation Engineering; B certificate holder by National Cadet Corps(Army); Hobby: Cooking, Yoga, Creative writing, drama; Math & Physics expert in Chegg India; Mentor for Data Science Students in Tapchief Project Information Sheet Analytics Projects Project: Sulvo Ad Price Predictor System – Sulvo (11/2016 – Till Date) Technology: Python, falcon, Google cloud platform, Node js, Tensor-flow, AWS, Redis, Neural Networks Details: The system has three major components – collector, real-time trainer, and predictor. It listens to the programmatic ad selling information for each impression through a collector and sends the predicted floor price for each impression to ad server through predictor. Real-time trainer builds the model on recently collected data and saves it, and predictor uses that model for prediction. Sulvo ad server is hosted in AWS cloud, and real-time trainer and predictor are hosted in Goggle cloud platform. The collector which is hosted in AWS and exposes a REST API to collect data receives information from the ad server and push it to a Redis message queue. On de-queue message from Redis server; the message is pushed to a Google Big Query instance. Real-time trainer which is hosted in a Google Compute Engine fetches the latest information from Big Query and builds a multi-layer CNN model using tensor-flow and save the model in Google Data Store. Predictor server which is a Falcon-based REST API is hosted on Google App Engine, receives prediction requests from Sulvo ad server and response back floor value calculated on basis stored data in the Data Store. The system is deep learning based, able to train the model with real-time data and also high performance, assured to send a response to ad server within 300ms and auto-scaled to serve 22 million prediction daily. Responsibilities: Building the product from scratch from data collector to ML model Project: RNN based Share Price Predictor – East India Security Ltd (8/2017 – Till Date) Technology: Python, Keras, AWS, Recurrent Neural Network, Error Feedback, Concurrent programming Details: This is a deep learning based sequence to sequence and sequence to label generation solution for predicting share price and also forecast if there is any lift or drop in next 100 bids with more accuracy. This is a real time system and system is completing its training in each iteration before the previous predicted sequence is completely used and prediction in lift and drop in next 100 bids has guaranteed accuracy of minimum 60 %. Responsibilities: Client handling, Model development and integrate it to client system Project: Topical Crawler for Web Harvesting - Abzooba (9/2015-12/2015) Technology: Python, Beautiful Soup, Graph traversal Algorithms Details: Project for a real estate client to automate manual reporting processes detailing data for High-Net-Worth clients. Developed a two step approach; gathered the URL with relevant information via a Tropical Crawling approach as well as extracting name, organisation and location from output URL via a name entity recognition algorithm. In terms of Topical Crawling, two sets of URL's were used for input, the first being seed URL's where crawling will begin and the second target URL's as reference URL's and a known subset of relevant URL's. Utilised Jaccard Distance as a similarity measure to compare content of new URL's with targets and select the most similar. This step was repeated to make the selected URL as the seed URL, which also considered all output URL's in previous steps and excluded those already selected. Reductions in search space were achieved by establishing a threshold in the similarity measure URL, where an average similarity is among the target URL's. A set of keywords was created, i.e. 'profile', 'biography', for the relevant URL string. Using a Name Entity Recognisation algorithm enabled the extraction of name, organisation and location in the selected output URL contents. Beautifulsoup library was used to parse the html pages and Stanford NER NLTK library for Name Entity recognition. Responsibilities: Problem Formulation, Design Solution Architecture, Mentor Implementation, Client Liaison Project: Vendor Recommendation using Cognitive Scale framework – Abzooba (2/2016-4/2016) Technology: Python, neo4j, angular js, Logistic Regression, Naive Bayes, Random Forest Details: Project to recommend the best vendors for a particular product ID using purchase order and invoice information. To address the problem this was divided into two sub-areas. The first area involved the prediction of probability of successful purchases. Three approaches were used in this calculation; a regression-based approach using Logistic Regression, a rule-based approach using Random Forest Algorithm and a Bayes-theory approach by Naive based classifier. The second phase involved grouping of data by VEN_ID and calculation of average probability of a successful purchase for each vendor. Vendors were then sorted according to this probability. Rule-based engines were then applied, such as vendor with minimum delivery time or vendor with maximum number of deliveries, and the most suitable vendor was recommended. Rules were prioritised by predicting aspects of user query using classification. Responsibilities: Problem Formulation, Design Solution Architecture, Mentor Implementation, Client Liaison Project: MCUBE Sentiment Analyzer - TCG Aviation Clients (7/2014 - 12/2014) Technology: Python, Java, MapReduce, Text Classification, Naive Bayes Algorithm, Elastic Search, Kibana, Mahout, Hive Details: Application text classification algorithm in brand reputation and promotion management, with input from various textual data in web, crm, mail server through crawler script or third party api and store data in HDFS. Classified sentences according to degree of sentiment (Positive / Negative / Neutral) as well as by brand dimension (Food / Staff / Entertainment); providing trend summaries / KPI information across individual brands to Data Brand Owners / Market Analysts. Selected naive bayes algorithm as text classifier following a literature survey, implemented via Mahout and visualised big data via Kibana; keeping the data in elastic search. Responsibilities: Requirement Analysis, Application Architecture Design, Database Schema Design, Unit Testing, Integrated Testing, Planning for Release Strategy, POC. Project: Nostradamus Approximation Framework - Inventory Estimation - Pubmatic (6/2012 - 12/2012) Technology: Java, Hadoop, MapReduce, Perl, MySql, Least Square Method, Correlation Analysis Details: Generic framework for estimation and forecasting of quantitative features from a huge historical data set; 100M orders per day and project to support 3 months of historical data. Proprietary implementation of Machine Learning methods enabled realisation of quantitative target features (No. of Impression, ECPM) as a linear regression of set of base features (Geo, Site, AdSize, Frequency). Applied linear regression to implement the Least Square method. Overall average accuracy improved by 60%+ and selected premium sites in smaller data sizes gave 80% accuracy. Responsibilities: Requirement Analysis, Application Architecture Design, Database Schema Design, Unit Testing, Integrated Testing, Planning for Release Strategy, POC. Project: METAL - Trading Time Risk Analysis Tool - Credit-Suisse (10/2009 - 2/2010) Technology: C# .net (WinForm), Market Data API, Position management System, WCF, Gemfire API, Oracle, Sql Server Details: Application Support project to modify and enhance an in-house position management, portfolio valuation, risk summary product. The application pulls market data through Reuters API and trade information from Fidesa, calculates various portfolio valuations and risk analysis parameters, then visualizes information to Traders. Responsibilities: Front / Back-end Design, Development, Unit Testing & Profiling, Beta Calculation, Average Volume Calculation, Industry Wise Index Decomposition, Exception Management, Risk Analysis Group Member (Imagine) of CSFB Prop-IT. Project: Price Optimization using Transaction Data - Abzooba (6/2015 – 8/2015) Technology: Python, pandas, Regression using Least Square, Quasi Newton Optimization Details: Project to examine what the sales price of a product needs to be in order to maximise profits. Maximising sales price may not result in maximised profit because the increment of sales prices can decrease sales volume / total profit. Other areas of impact can be cash back amount and product purchase price as well as the limitations of data incompleteness, particularly in respect of zero information on unsuccessful product sales. Analysis is completed on transaction data provided by the leading Indian Ecommerce Company. Sales volume is realised as linear regression of sales price, total cash back and buy price using least square method, then expressed as profit as a function of SP, TCB, and BP. Using mathematical optimisation with constrain in all parameters values enables the identification of optimal values of the parameters which maximise profit. We used Quasi-Newton method for this optimization problem. Responsibilities: Problem Formulation, Design Solution Architecture, Mentor Implementation, Client Liaison Project: Trend Analysis of Issue Register - Credit-Suisse (8/2008 - 12/2008) Technology: ASP/C# .net, Win32 Module, XHTML, AJAX, Moving Average Analysis, Control Charts, Google Visualisation Tools. Details: Operational process improvement initiative to resolve various application, server and network related issues. Analysed data from MS Excel and derived statistical parameters indicating device health / criticality issue trends and signal to noise ratio in alerts. Published webpage containing GUI interface of trend analysis. Responsibilities: Proposing Statistical Parameter for Device’s Criticality, Module Development. Additional Projects Project: Real Time Latency Monitoring in High Frequency Trading - Credit-Suisse (9/2008 - 10/2009) Technology: C++, Shell Scripting, Unix Kernel Programming, Tibco EMS, Socat, FIX protocol, Corvil, Memory map file, ARM, Kronos, My-ARM Details: Developed a virtual order management system and simulated to a high transaction environment. Instrumented by ARM, MY-ARM and Kronos API and compared performance characteristics, such as accuracy, instrumentation overload in execution time, memory and processor usage. Parsed log file in trading time, then designed and implemented a distributed, divide and conquer architecture for Real-time Log-file Parsing Framework. Data of potential interest is transmitted to a central server through a robust and highly tuneable transport layer. Completed heavy weight parsing in the central server separated to the production host. Divided order latency into two phases (inside the box / outside the box) to enable further analysis; inside the box latency measured by Kernel program / Outside the box latency collected by Corvil API using WSDL. Responsibilities: Design, Development, Performance Testing and Production Release of Socat base shell script, C++ base log Development. Project: SMILE - Online Loyalty Reward Program - FairTech (11/2010 - 12/2011) Technology: C#/ASP .net, SQL Server, Web based API (NVP/SOAP/AJAX), PayPal Adaptive Payment API, CheckOut Flow, NHibernate, WCF Details: Patented a differential reward point strategy which created equilibrium among the all payment options for merchants. Designed a web application distributed in merchants and LRP vendor website which was scalable and enhanced LRP infrastructure. Merchants develop SLA with the LRP vendor, who in turn embeds a module in the Merchant's website payment gateway or the LRP vendor will build the website in none in place. Consumers registered on the LRP website issue payment to the Merchant website and SLA between the Merchant and LRP consumers receive loyalty reward points, with the transaction recorded on the LRP database / website. Merchant can login into LRP website and buy the reward points for vendor also can avail the Customer Relationship Management (CRM) facility in the LRP website. Consumer can login in to the LRP website to see transaction details and covert reward point to money / vice-versa.. Merchant and Consumer can build a network via the website and view granted transactions. Merchant can also generate gift coupons that consumers can request. A special donation page in the LRP website allows donation candidates to register details for approval by LRP vendor. Responsibilities: Requirement Analysis, Application Architecture / Database Schema Design, Team Leadership (3 direct reports), Unit Testing, Integrated Testing, Release Strategy Planning. Project: TRIANA Dashboard - Credit-Suisse (10/2008 - 9/2009) Technology: C# .net (WPF), WSDL client, Nikson, SevOne, Ganglia, Oracle, Corvil Details: Project to develop a Trader’s Infrastructure Application Network Monitoring Tool; integrating network, server, application alerts in a common dash board for all devices under a particular prop trader. Front end is a grid view of WPF application. Back ends are different adapters those run in parallel threads. For network alerts we use SevOne API through WSDL client. For server alerts we use Ganglia API through telnet client. For application alerts it gets data from web server which stores the application alert through Log-file parsing project. When a device name is entitled in dash board connected device names also come through firing query to configuration database. It is an in-house enterprise product of Credit-Suisse. Responsibilities: Design, Develop, Unit Testing, Profiling Data Adapter parts (Ganglia, SevOne, Application alerts), Front End Development a explorer view bind to a database through Xml Document