Tushar Gupta

Machine Learning Engineer

Bengaluru, Karnataka, India2 yrs 10 mos experience

Highly StableAI Enabled

Key Highlights

Designed an NL-to-SQL system improving analysis time by 40%.
Awarded Samsung Spot Award for outstanding contribution.
Hands-on experience with cutting-edge AI technologies.

Stackforce AI infers this person is a Machine Learning Engineer with expertise in AI-driven solutions and web development.

Contact

Skills

Core Skills

Natural Language Processing (nlp)Large Language Models (llm)React.jsJavascript

Other Skills

AI FairnessArtificial Intelligence (AI)BashCCSSDSAData AnalysisData StructuresDecision Tree ClassifierDeep LearningExploratory Data AnalysisHTMLK-Neighbors ClassifierLinuxLogistic Regression

About

AI Developer at Samsung with 2 and half years of experience spanning software development, system performance optimization, and now cutting-edge work in AI and LLMs. 🌟 Currently building AI Agents from scratch to automate Data Analyst workflows — from natural language interaction to SQL generation and real-time data retrieval using Samsung’s DCS data portal. 🔍 Previously optimized a large-scale C++ testing framework, improving execution time by 40% using parallel execution and memory fixes. Awarded the Samsung Spot Award for outstanding contribution. 🚀 Tech I'm hands-on with: Python, LLMs, RAG (Retrieval-Augmented Generation), Streamlit, VectorDBs, SQL, and C++ (legacy projects). I’m now exploring AI-focused or software developer roles where I can scale these learnings and contribute to high-impact products. Let’s connect!

Experience

2 yrs 10 mos

Total Experience

2 yrs 10 mos

Average Tenure

2 yrs 10 mos

Current Experience

Samsung r&d institute india - bangalore

Machine Learning Engineer

Aug 2023 – Present · 2 yrs 10 mos · Bengaluru, Karnataka, India · On-site

Designed and optimized a Large Language Model–driven Natural Language to SQL (NL-to-SQL) system for complex query understanding and structured data retrieval.
Performed LoRA-based parameter-efficient fine-tuning on proprietary NL–SQL pairs, directly modifying model weights to improve semantic-to-syntax mapping accuracy.
Further enhanced model robustness through a Reinforcement Learning from Human Feedback (GPRO) loop that leveraged user corrections for continuous retraining — demonstrating deep understanding of LLM fine-tuning, adaptation, and performance optimization.
Engineered an NL-to-SQL pipeline with query decomposition, translating complex natural language queries into sub-queries and merging their outputs into final executable SQL commands.
Integrated the pipeline into a Full-Stack AI Assistant built using Python, LangChain, Llama, RAG, Vector Databases, and SQL, reducing analysis time by 40% and achieving 97% mapping accuracy on sub-query translation.

Natural Language Processing (NLP)Large Language Models (LLM)SQLPythonReinforcement LearningData Analysis

Feevin media

Web Developer

Dec 2022 – Mar 2023 · 3 mos · Remote

Developed a React-based NEWS WEB Application using class-based components in React.
Made the admin panels for the multiple Clients with all kinds of features like updating the blogs, deleting/adding blogs and content on the website. Etc.
The Tech Stack includes ReactJs, HTML, CSS, Java Script(For Frontend), PHP(for Backend).

React.jsJavaScriptPHPHTMLCSS

Smartknower

Machine Learning Intern

Nov 2020 – Dec 2020 · 1 mo · Remote

Raw dataset provided by the team learned about EDA(Exploratory Data Analysis) while cleaning data, and using various ML models while analyzing it like LogisticRegression, KNeighborsClassifier, DecisionTreeClassifier, RandomForestClassifier, VotingClassifier.
Project demand is to analyze the typos made by men and women while using twitter.

Exploratory Data AnalysisLogistic RegressionK-Neighbors ClassifierDecision Tree ClassifierRandom Forest Classifier