Engineer, CTO - part of 4 startups during key growth phases

Engineering Manager @ Helping with the Brand Advertising strategy at Facebook From September 2015 to Present (4 months) London, United KingdomCTO @ Responsible for the technical strategy, roadmap, planning and executing the vision of the company. Duties include technology P&L, hiring and training, evaluating/buying technologies and seeking mutually beneficial partnerships with other companies. From April 2011 to August 2015 (4 years 5 months) London, United KingdomSenior Developer @ - Engineer #3 to start at Skimlinks. Built Skimwords - a system for auto-linking phrases on publisher websites based on phrase performance, profile of the user visiting the page and the content of the page. This involved building a recommendation engine (k-NN), a webpage classifier, a persistent trie and a rules engine. When I left, Skimwords was responsible for bringing in over 40% of the company revenue. - Performed technical due diligence and recommended the acquisition of Atmalinks. - Built a system for crawling, parsing, indexing and search to power the public API and Skimkit. This involved extending the Python XML sax parser to parse huge (30+ GB) XML files, a Message queue to communicate and pass data between the various components and writing a de-duplication and re-ranking plugin for Solr. From February 2010 to April 2011 (1 year 3 months) London, United KingdomSr. Product Engineer @ - Built a Semantic Ads framework to automatically extract information from an unstructured classifieds listing. Also built the corresponding search component to output context sensitive ads. This involved extending the NLTK POS-Tagger, writing a custom chunker and grammar for Hinglish, a naïve Bayes classifier, a kd-tree implementation and a concept extractor for a webpage. - Implemented a common indexing and search framework using Solr to get all the search verticals (Book search, Media share, QnA, etc) under one platform. This involved writing custom Solr plugins for de-duplication, ranking and grouping for each vertical. - Successfully led and managed two teams of size 4 and 12 for the above projects. This includes project management for the entire life cycle of the products. From September 2008 to May 2009 (9 months) Mumbai Area, IndiaCo-founder @ Created and exited a 2-person startup in the area of localized classified search. Was responsible for design and implementation of the whole system. This involved building a crawler, parser, NER extractor, a custom chunker, a rules engine and a nearest neighbour search (NNS). From January 2008 to August 2008 (8 months) Mumbai Area, IndiaResearch Assistant @ Implemented a Transliteration Engine to transliterate between 13 Indian languages and English. Also implemented a document classifier and a summarisation engine. From February 2007 to November 2007 (10 months) Mumbai Area, IndiaSoftware Engineer @ Engineer #8. Built a Closed Crawler, Indexer, Packaging system, Rewriting engine, Search system and an HTML parser for the Desktop and Windows Mobile Devices. This involved writing a Priority Queue, a State Machine to parse/rewrite HTML, a managed wrapper over Tidy HTML parser, modifying Lucene to work on Windows Compact .NET Framework and writing custom Windows UI elements. From May 2005 to January 2007 (1 year 9 months) Mumbai Area, India

BTech, Computer Science @ Indian Institute of Technology, Bombay From 1999 to 2003 Pravin Paratey is skilled in: Natural Language Processing, Machine Learning, Big Data, CTO, Python, Linux, Scalability, Algorithms, Advertising, Online Advertising, Data Mining, Lucene, Search algorithms, Hadoop, Natural Language Understanding