Devansh Devansh Email & Phone Number

New York, Ny, Us

New York, Ny, Us

New York City Metropolitan Area

Helping companies integrate Artificial Intelligence, Machine Learning, and Data Science into their operations to improve KPIs. Some notable projects include- Deepfake Detection- Extremely low-costs, generalizable, Deepfake detection solution that detected Deepfakes, AI Altertered Images, and other sources of misinformation with 85% score and low costs. Helped a Financial Services Firm rearchitect their Retrieval Augmented Generation System to reduce hallucinations, strengthen guardrails, and improve their source ranking. Reduced the memory and computational costs of an e-commerce firm by 20% by improving their chunking and semantic retrieval strategy. Worked with Legal Tech Firm to build an LLM-enabled AI Agent for relationship visualization, document analysis, case sourcing, drafting and other relationships. Reduced Lawyer Worktime by 50-75%. skills- presales, client development, sales engineering, managing AI teams, ML Engineering, Consulting

New York City Metropolitan Area

Helping Clientell with Machine Learning tasks when required.

New York City Metropolitan Area

 One of the 3 Cofounders. Deployed Machine Learning Pipelines for Clientell, a no-code B2B SAAS platform helping organizations streamline their revenue operations. Solutions were responsible for improving client sales forecast accuracy to 95% and improving quota attainment by 23%. Notable Projects include- Text Analysis Bot-o Led the development of an information mining bot from emails using Natural Language Processing, Transformers, Regular Expressions, Semantic Matching and Named-entity Recognition for extracting key information from emails, text conversations, and other CRM platforms. Work was done using Python and Cloud Computing.o The Bot led to 15% reduction in missed meetings and 5% improvement in productivity for service representatives.o Quantified deal engagement using Sentiment Analysis, Large Language Models, and HuggingFace Transformers. Automated Machine Learning Pipeline-o Automated an end-end Machine Learning Pipeline- from data cleaning to model retraining and deployment.o Integrated various Data Augmentation, Imputation, and Processing Processes- while comparing various standard models- Random Forests, XGBoost, GBT, Neural Networks, SVMs, KNNs etc on a variety of metrics.o Saved 200 hours of dev time monthly in searching through over 50,000 training protocols and configurations. Financial Planning Feature-o The tool allowed clients to forecast their sales, identify sources of churn, and estimate future revenues. This enabled better planning for clients.o Implemented time series forecasting by leveraging genetic algorithms, Bayesian networks, and Recurrent Neural Networks (LSTMs). Improved Forecasting error by 83.333% using Neural Architecture Search and Deep Learning.o Implemented Bayesian inference (Bayesian Neural Networks and Belief Networks) to create simulations with incomplete information. Fine-tuned simulations with statistical analysis. Enabled clients to test different policies.

New York City Metropolitan Area

Machine Learning- Built the AI pipelines for Clientell, to raise the initial funding rounds and onboarding the first clients.

United States

Creating in-depth technical content covering ideas on Software Engineering, Machine Learning, Artificial Intelligence, Tech Leadership, and more. Built and managed a community of AI Researchers, Academics, Engineers, and Managers and other Tech Professionals Some numbers for you (March 2023)Monthly Views- 1M+Email Newsletter Subscribers- 30K+Content consumed in over 150 Countries

Miami, Florida, United States

Created API that generated automated Machine Learning Reports for users. Developers just had to call the dataset and the pipeline automatically generates the reports, handling all the parts. Built the pipeline end-end. Integrated data encoding (for categorical variables), multiple data imputation policies, multiple error metrics and models, feature importance analysis, EDA and much more. The API automated routine tasks for Machine Learning Engineers in KlearNow, significantly saving developer hours. The API+ scripts have now been deployed for their uses.

Memphis, Tennessee, United States

Evaluated Supply Chains for potential areas of risk. Developed Forecasting tools to predict things based on behavior. Scrape Information to evaluate risk in different geopolitical situations1) Analyzed supplier data to calculate supplier risk. Used time series, sentiment analysis, genetic algorithms, Bayesian networks and Recurrent Neural Networks(LSTMs). Reduced Forecasting error by 500% using Neural Architecture Search. 2) Implemented Bayesian inference (Bayesian Neural Networks and Belief Networks) to create simulations with incomplete/partial information. Fine-tuned simulations with statistical analysis. LinkedIn Recommendation available.Presented all of these to CEO and other stakeholders along with suggestions for proceeding

Rochester, New York, United States

Student Researcher assisting Dr. Matthew Wright. Working with him into looking into deepfake detection techniques to identify fake images in a reliable way

Mumbai, Maharashtra, India

Topics: Python, Google Analytics Platform, DataDiscovery, CloudSQL, BigQuery, Software Engineering1)Developed a framework for identifying potentially high-value customers. Ran queries on the cloud, built software tools using Python and GCP to enable the Bank to retrieve valuable information. LinkedIn Recommendation Available2)Analyzed user behavior on the website to identify customers that were likely conversions. Used that to Identify most important predictors of conversions, saving the bank many hours of follow up on low-value customers. Used this information to classify customers with an average of 95% accuracy.

Baltimore, Maryland, United States

My work involves working on various public health datasets. I perform tasks such as EDA (exploratory data analysis), Data Modeling, Hyper Parameter Tuning, and Ranking Features for different datasets and procedures.Topics: Python, Machine Learning, JS, Data Visualization, HTML, Software Development Worked with Dr. Shivam Gupta and MD Zabir Hassan at the School of Public Health. Analyzed data collection policies of State Governments and their impact on health system and public health. Generated reports on the relationships between different data. Tested various supervised algorithms, engineered features and analyzed the clustering of combinations in higher dimensional spaces. Achieved 97% accuracy with k-fold cross validation of 10. Compiled different types of errors, learning and accuracy curves to give comprehensive perspective. Derived most important aspects of health systems (on public health). This allowed the state government to invest more in these areas. Also, Saved government money by identifying the programs and ventures with low returns on investment. Work used to inform State Policy of a state with population of 204.2 million people.

Rochester, New York Area

Demo: https://people.rit.edu/dl1683/WongTopics: Python, Machine Learning, JS, Data Visualization, HTML, Scripting, Software Development Demo Hosted here. Built for Professor Tony Wong for modeling and viewing climate and environment changes based on different simulation variables. Product was an interactive visualization tool that allowed him to vary different parameters to compare the difference in results using JavaScript (D3) and full web stack. Simplified comparing any combination of 48 possible simulations to a few clicks. Automated visualizing over all possible configurations (search space grows exponentially as we try to add another model into configuration. Possibilities with 1 config->48, w/ 2->1128, w/ 3->17296 so on).

Tutor at the office of K12 outreach at RIT. Help children with homework, assignments, and difficult concepts

Rochester, New York Area

LeverEdge is the first collective bargaining group for student loans and has saved students over $20M in interest and fees to-date. As a Growth Marketing Manager at RIT, I'm responsible for user acquisition and retention through a variety of off-line and online channels

Rochester, New York Area

Pune, Maharashtra, India

Finance Terms: https://people.rit.edu/dl1683/Docs/Patent Devansh.pdfTopics: Python, Machine Learning, Shell Scripting, Java, Software Development, Linux Collaborated with the Dean of Technology on multiple projects including fingerprint recreation, intelligent systems and Parkinson’s Disease detection. Worked on different OS to test different projects and scripts. Automated low-level tasks for doctoral students and professors with their projects. Developed scripts and tools for allowing instant use. Developed software to assist with multiple tasks, saving hours of researcher hours. Developed proprietary, low cost, Parkinson’s Detection Algorithm using voice samples with 92 percent accuracy and low cost. Commercialized for INR 100,000 + Stake in every implementation. Financial terms here.

In this internship, I had the privilege of gaining exclusive experience as a mentee to Phani Bhusahn, the founder of Anant Computing. With his expertise and insight, I learned to develop, optimize, and apply a number of applications in JavaScript. As well, I acquired hands-on experience in Lean Development methodology and application. Results:Topics: JavaScript, Android Dev, Application development methodologies Created the scripts to integrate various functionalities into different phone applications Developed and tested low level applications for diverse applications in JavaScript for the compatibility with a custom software development kit.  Built GUI parts, and integrated multiple (natural)languages.  Worked with various platforms to evaluate benefits of integration.

Pune, Maharashtra, India

This was my first time working with Dr. Abhyankar, the school’s Dean of Technology. I assisted his team on various projects, where I would often do handle tasks like automation of tasks or testing.

Bsc - Bachelor Of Science, Computational Mathematics With Minor In Computer Science And Immersion In Gender Studies

High School Diploma