Tushar Makkar

Making intelligent robots for one of the most challenging and labor-intensive sector.

Leading AI platform at Eightfold. Spearheading Knowledge Graph's & LLM's based initiatives across Eightfold.

Lead Data Science team in India. Spearheaded Document AI, Talent Management AI & Job Intelligence Engine AI's initiatives across Eightfold. Mentoring a team of 12+ highly talented & motivated full stack software engineers, Machine learning engineers and senior engineers.

Solving critical business problems related to logistics industry and building marketplaces. Helped solve problems related to recommendation engine and pricing engine.

Helping in solving data science problems in real estate industry. Worked on building geo location de-duper and pricing models.

Converting unstructured data to knowledge using Machine Learning. Building Resume Parsing using deep learning techniques from ground up.

Worked on making custom content manager using Jupyter. Modified source code of open source Jupyter, Jupyter Lab and Jupyter Hub to achieve the same.

Solving employment – finding the right person for the right role at the right time. Senior Engineer in Eightfold.ai's Data Science Team. Building state of the art deep learning models for matching & resume parsing.

Formulated and solved following problems: - Abbreviation Expander: Improved engine's performance by integrating abbreviations semantical information. - Spell Checker: Internationalised and Fast spell checker library. - Para-BERT: Contextual BERT embeddings for different entities in Eightfold's ecosystem.

Made algorithms for making Freight Pricing Engine the universal go-to market rate. National Freight Index (NFI) - NFI is a barometer of the road freight spot market, offering an aggregated picture of both, live rates and historical trends of spot prices across 150 different combinations. NFI can cater to use cases for logistics decision makers, researchers, sales force of trucking related businesses (OEMs, NBFCs), researchers, academicians and students. The underlying data used for computing the indices is based on Rivigo’s marketplace transactions and machine learning and economics principles powered algorithms that use millions of data points. Relay as a Service Simulation Platform: Built a scalable and extensible simulation platform from scratch incorporating details of Rivigo's Relay algorithm. Incorporated various algorithms like: Pilot allocation engine, Vehicle planning engine etc. in it to make it as close as real operational system. Added machine learning based models for detention prediction and wait for load computation in the simulation platform. Network Equations: Made equations for running the network. It consists of 6 equations and calculates the gross margin per hour which will be incurred if the given demand is accepted. It is calculated using recursive formulation of incorporating current trip margins and future trip margins. It also incorporates the factors like rejection cost of given demand, wait for load for next load at destination, network balance/imbalance happening due to acceptance of current demand and future demand, detention at client warehouse and dry run in current and future trips. It showed 10% growth in Gross Margin/hour via simulation engine. Pilot Resource Planning (PRP): Made stochastic gradient descent based algorithm for finding optimal number of pilots at a given pitstop and testing the results using Simulation Platform. This showed around 16% increase in truck utility and 8% increase in pilot utility.

Working in Rivigo Freight / Vyom Team as senior software engineer in Data Science team. Freight (Vyom) is a unique network of all stakeholders in the trucking industry. Freight (Vyom) defines the next generation trucking industry with definite predictions, high levels of efficiency and simple automated operations. I am a member of Data Science team and lead the Pricing Engine. Pricing Engine enables Vyom to present best route price for transporters and supply side to ensure better return load efficiencies and more business for fleet owners/brokers using machine learning models. Worked on building and architecting World's First Truck Freight Exchange.

* Developed Library Suggestion Engine (Patent Granted), a product which analyses code and yield library substitution opportunities. * Worked on developing IWS (Insights Web services), a web services platform for enabling development of products faster. * Played crucial role in hiring and setting up processes. * Mentored Interns and New joinees.

Graduated from CodeNation University 2015 (CNU'15) / Trilogy University 2015 (TU'15). Researched, developed and productionalized Library Suggestion Engine.

Graduated from CodeNation University 2015 (CNU'15) / Trilogy University 2015 (TU'15). Formulated an innovative and intelligent code analyser for JAVA using NLP, AI and ML.

Founding engineer of the team. Worked on designing and establishing new true P2P cryptocurrency which will give better security, privacy, anonymity and transparency. Made a Biometric User Authentication system using Handwriting recognition (ML Based) and OCR techniques.

Worked on Word Sense Disambiguation problem in Ray Kurzweil’s Machine Intelligence Group along with Shailesh Kumar. It is an open problem of natural language processing and ontology, which governs the process of identifying which sense of a word (i.e. meaning) is used in a sentence, when the word has multiple meanings.

Worked on implementation and design of Least Cost Fulfillment Algorithm which aimed at systemizing the order fulfillment process of The Supplies Group. Formulated Genetic Algorithm Based approach for reducing the time taken by the exhaustive search algorithm.