Rameshwar Garg

Data Scientist

New York, New York, United States3 yrs 7 mos experience

Most Likely To SwitchAI ML Practitioner

Key Highlights

640x improvement in memory efficiency for LLMs.
Led a team to develop an innovative RAG pipeline.
Achieved 36% improvement in monsoon forecasts.

Stackforce AI infers this person is a Data Science and AI professional with a strong focus on Generative AI and Machine Learning.

Contact

Skills

Core Skills

Machine LearningData ScienceGenerative AiLarge Language Models (llm)Application DevelopmentData EngineeringComputer VisionDeep LearningFeature Engineering

Other Skills

Foundation ModelsArtificial Intelligence (AI)Applied ResearchHyperparameter OptimizationResearch SkillsLanguage ProcessingRetrieval-Augmented Generation (RAG)Amazon Web Services (AWS)Python (Programming Language)Data AnalyticsBayesian OptimizationUnstructured DataMLOpsConvolutional Neural Networks (CNN)Predictive Modeling

About

I am a performance-driven individual with an eagerness to accept challenges while learning and gaining experiences from them. Pursuing my M.S. in Data Science at Columbia University has given me the opportunity to revel in my passion for Machine Learning, Data Science, Deep Learning & Generative AI. I am now working as a Data Scientist and and hoping to make a discernible difference in the world. Machine Learning, Deep Learning & Gen AI are my primary areas of interest as far as research is concerned. On a personal note, I am very enthusiastic about cricket and drumming. Being the outgoing person that I am, I have also worked with International NGOs to build and provide housing facilities to the underprivileged. I enjoy the role of teaching, and actively participate with NGOs to teach English, Mathematics and Life Skills to government school children. At wok, I have also led various efforts to drive engagement and improve life in the workplace. I enjoy working with teams as it helps me improve as an individual and as a leader. I have an optimistic outlook and believe that there is always a chance to learn and grow from every situation. I strive to keep learning, and, in the words of Socrates, believe that Education is the kindling of a flame, not the filling of a vessel.

Experience

3 yrs 7 mos

Total Experience

1 yr 2 mos

Average Tenure

1 yr 5 mos

Current Experience

Capital one

Senior Data Scientist

Jan 2025 – Present · 1 yr 5 mos · New York, New York, United States

* Working with Foundation Models for Customer Data in the Applied AI Research Org, to improve the quality of prediction for downstream Machine Learning tasks.

Foundation ModelsArtificial Intelligence (AI)Generative AIApplied ResearchMachine LearningData Science+2

Columbia university

AI Research Assistant

Sep 2024 – Jan 2025 · 4 mos · New York, New York, United States

Optimized Large Language Models for tabular data by extracting key features, resulting in a 640x improvement in memory efficiency.
Prepared a Tabular Understanding Dataset, which enables LLMs to understand table semantics.
Fine tuned a long context LLM with the above dataset, to improve its performance on tabular tasks.

Research SkillsGenerative AILarge Language Models (LLM)Language Processing

Johnson & johnson

Applied AI Capstone Intern

Sep 2024 – Dec 2024 · 3 mos · New York, United States

Orchestrated a 5-member team to create dossiers with an LLM based Agent and Retrieval Augmented Generation (RAG) in PyTorch.
Developed a comprehensive RAG pipeline with thorough evaluation methods such as statistical evaluation and LLM-as-a-Judge based evaluation, with a Human-in-the-Loop to mitigate hallucinations.
Presented a poster based on work, at an AI Event hosted at Columbia University.

Generative AILarge Language Models (LLM)Language ProcessingRetrieval-Augmented Generation (RAG)

Mavenir

Data Science Intern

Jun 2024 – Aug 2024 · 2 mos · Dallas, Texas, United States

Enhanced performance of random forests for handover failure rate prediction through a custom weighted regression approach by 9%.
Implemented bayesian optimization, optuna, successive halving for hyperparameter optimization thus increasing search space by 150x.

Deep LearningArtificial Intelligence (AI)Data EngineeringHyperparameter OptimizationAmazon Web Services (AWS)Python (Programming Language)+6

Columbia university

Research Assistant

Jan 2024 – May 2024 · 4 mos · New York, New York, United States

Improved Indian monsoon sub-seasonal forecasts through deep learning based non-linear calibration by 36% using Python & Keras.
Leveraged convolutional neural networks for deterministic regression & probabilistic classification for bias correction on 3 datasets.

Deep LearningConvolutional Neural Networks (CNN)Python (Programming Language)Machine LearningPredictive ModelingData Science+2

Cisco

2 roles

Software Engineer

Aug 2021 – Jun 2023 · 1 yr 10 mos · Bengaluru, Karnataka, India · Hybrid

Created a predictive model to forecast all metrics associated with network devices, with an accuracy of 83% leveraging Python.
Implemented a scalable, big-data pipeline to classify customer contracts using Python & AWS, with a processing time of 7 seconds.
Collaborated on automating the software upgrade of network devices using AWS & Spring-Boot, reducing a 6-month task to 1 hour.
Optimized a cloud platform to provide Day 0 to Day N support & performance analysis for network devices, with AWS & Spring-Boot.
Integrated Cisco's cloud platform with ServiceNow, enabling bidirectional communication for customers to raise tickets in 5 minutes.

Application DevelopmentSpring BootData EngineeringServiceNowPostgreSQLGit+12

Software Intern

Feb 2021 – Jul 2021 · 5 mos · Bengaluru, Karnataka, India · Hybrid

Developed an ML workbench to forecast network KPIs and deliver insights about network capacity with Streamlit & FB’ prophet.
Presented varied forecasts, confidence intervals, robust back-testing solutions, split-screens & improved forecasting accuracy by 12%.

BitbucketAnacondaPostgreSQLGitTime Series AnalysisProphet+11

Samsung r&d institute india

Summer Research Intern

May 2020 – Jul 2020 · 2 mos · Bengaluru, Karnataka · Remote

Conducted research to remove flash glare defects from images in real time, using computer vision & image processing in Python.
Analyzed the capability to maintain color & texture, and studied the time efficiency of 17 different statistical & deep learning models.

Deep LearningImage ProcessingStatistical ModelingComputer VisionComputer ScienceData Structures

Florida international university

Summer Research Intern

Jun 2019 – Jul 2019 · 1 mo · Miami, Florida, United States · On-site

Pioneered 2 feature selection algorithms combining filter and wrapper approaches, to pick high relevance & low redundancy features.
Validated these models on 12 datasets with pandas, numpy & scikit-learn whilst supervised by Dr. S.S. Iyengar & Dr. Thejas G.S.
Improved performance of ML classifiers by around 9% for real-time data and approximately 25% for benchmark data.

Feature SelectionResearch SkillsFeature EngineeringPython (Programming Language)Machine LearningData Science+2