Vivek Gupta
Verified Expert in Engineering
Data Scientist, Machine Learning Engineer and Developer
Bangalore, India
Toptal member since July 7, 2026
Vivek brings extensive experience in applying machine learning (ML) and generative AI (GenAI) across diverse fields and complex challenges. He has built end-to-end agentic workflows and RAG pipelines. He is proficient at translating business goals into data products and at architecting the entire pipeline from end to end. Vivek has led teams and shipped ML systems from concept to production across fintech, eCommerce, and consulting in both startup and enterprise environments
Portfolio
Experience
- Data Science - 10 years
- Artificial Intelligence (AI) - 10 years
- Python - 10 years
- SQL - 6 years
- LangGraph - 2 years
- Retrieval-augmented Generation (RAG) - 2 years
- Multi-agent Orchestration - 2 years
- Reinforcement Learning - 1 year
Work Experience
MTS-1 Machine Learning Engineer
PayPal
- Led architecture and 0-to-1 delivery of the KYC ops assistant—production agentic AI service (LangGraph, CrewAI) on a fintech compliance surface. Defined the orchestration layer for models, tools, memory, and decision-making. Doubled KYC conversions.
- Designed the RAG pipeline (chunking, embeddings, vector retrieval, hybrid search and re-ranking, citation enforcement) and integrated multiple MCP tools for workflow automation—exposing internal services as guarded, schema-validated agent tools.
- Built an internal LLM evaluation and observability platform used across DS/ML teams for multi-agent system evaluations, LLM-as-judge, RAGAS, guardrails, and offline and online evaluation tied to quality, latency, and hallucination.
- Fine-tuned LLMs as digital agents (Venmo) using SFT and RLHF (PPO, GRPO) on multi-GPU clusters—owned training, evaluation, monitoring, and production rollout.
Lead Data Scientist
Cars24
- Managed a team of 5 data scientists—hired, set roadmap and priorities, ran code/model reviews, mentored on system design and production reliability.
- Owned the dynamic pricing engine end-to-end—4% gross-margin lift at scale with SHAP explainability and A/B testing as source of truth; designed around latency, cost, and conversion-velocity tradeoffs.
- Developed an in-house job allocation model using the Linear Sum Assignment algorithm, achieving 8% reduction in travel distances for car inspection mechanics.
Senior Associate, Data Science
Bain & Company
- Embedded technical advisor for Fortune-500 retail and telecom leadership—scoped workstreams and translated model behavior into executive decisions.
- Built demand-forecasting and price-optimization models (LightGBM, Hyperopt)—5% reduction in promotion spend.
- Built NLP pipelines (topic modeling, sentiment) for a telecom client's churn strategy.
Lead Data Scientist
IND Money
- Built and shipped the mutual fund recommendation and ranking engine (XGBoost and similarity ranking)—a production fintech personalization surface serving end users daily.
- Built the investment portfolio analysis engine that provided comprehensive analysis of users' portfolio performance compared to benchmarks, return and risk parameters, diversification, investment patterns, and hidden expenses.
- Established the data-science function from scratch as the 10th employee, owning the end-to-end lifecycle.
- Designed an AI-driven SMS transaction-parsing system (NER and supervised text classification)—large-scale data enrichment, classification, and correlation across transaction data with a labeling pipeline that continuously improved precision/recall.
Data Scientist
Cogneau Systems
- Optimized a distribution-center network for a Malaysian logistics customer using DBSCAN—reducing operations cost by 30%.
- Delivery time prediction for a shipping client based on regression models.
- Built all projects from scratch, working as the only data scientist in the team.
Data Scientist
Simpli5d Technologies
- Built a real-time bot-detection classifier (Random Forest on Spark/PySpark) processing 10+ million user sessions per day across 120+ publishers.
- Designed the Spark feature pipeline and distributed-data infrastructure—high-throughput, low-latency event processing.
- Delivered the project single-handedly, working as the only data scientist in the team.
Decision Scientist
Mu Sigma
- Analyzed search-advertising KPIs across Asia Pacific for Bing Ads.
- Created a strategy for drug distribution for a pharmaceutical company.
- Applied skills such as statistical modeling, linear regression, and logistic regression during the training.
Project History
Multi-agent AI Chat & Dashboard with RAG
I designed and implemented the RAG pipeline (chunking, embeddings, vector retrieval, hybrid search and re-ranking, and citation enforcement) and integrated multiple MCP tools for workflow automation—exposing internal services as guarded, schema-validated agent tools.
Fine-tuning Open-source LLMs – Venmo Assistant
Multi-agent RAG System Evaluation
Dynamic Pricing ML System for Selling Used Cars – Cars24
Price Optimization During Sale Period for a Global Retail Giant
Portfolio Analysis and Recommendation
Education
Bachelor's Degree in Biotechnological Engineering
Indian Institute of Technology Kanpur - Kanpur, India
Certifications
Semantic Caching for AI Agents
DeepLearning.AI
Claude Certified Architect – Foundations
Anthropic
How Diffusion Models Work
DeepLearning.AI
Long-Term Agentic Memory With LangGraph
DeepLearning.AI
Generative AI: Working with Large Language Models
LinkedIn Learning
Improving Deep Neural Networks: Hyperparameter Tuning, Regularization and Optimization
Coursera
Structuring Machine Learning Projects
Coursera
Skills
Libraries/APIs
XGBoost, PyTorch, Scikit-learn, Pandas, OpenAI API, REST APIs, NumPy, Claude API, Playwright, CatBoost, PySpark, vLLM, DeepSpeed, Hugging Face Transformers, Flask API
Tools
Git, Jupyter, Claude Code, Claude, ChatGPT, Claude Agent SDK, BigQuery, Azure OpenAI Service, Microsoft Excel
Languages
Python, SQL, JavaScript, TypeScript, R
Frameworks
LangGraph, LightGBM, Agentic Frameworks, DSPy, Spark
Paradigms
Model Context Protocol (MCP), Automation
Storage
PostgreSQL, Data Pipelines
Platforms
Azure, LangSmith, CrewAI, Amazon Web Services (AWS), Google Cloud Platform (GCP), Docker, Kubernetes, Linux, Visual Studio Code (VS Code), Vertex AI
Industry Expertise
Automotive
Other
Retrieval-augmented Generation (RAG), Large Language Models (LLMs), LangChain, Artificial Intelligence (AI), Agentic AI, Agentic RAG Systems, Dynamic Pricing, Statistics, Machine Learning, Random Forests, Model Evaluation, Stakeholder Management, Agentic AI Systems, AI Agents, LLM Integration, Debugging, Decision Trees, Chatbots, Vector Databases, Generative Artificial Intelligence (GenAI), Prompt Engineering, Data Science, Data Scientist, Multi-agent Orchestration, Machine Learning Operations (MLOps), Data Analysis, Data Engineering, RAG Architecture, Weaviate, Workflow Automation & System Integration, Workflow Automation, AI Integration, Agentic Workflow Design, APIs, Open-source LLMs, AI Assistants, Cursor AI, Multi-agent Systems, Feature Engineering, Linear Regression, Communication, Model Development, RAG Pipelines, RAG Systems, AI Modeling, Model Tuning, Documentation, API Integration, Datasets, Anthropic, AI Chatbots, Conversational AI, Data Analytics, Predictive Modeling, AI Agent Orchestration, Orchestration, Back-end, Back-end Development, Metrics, Time Series, Dashboards, Architecture, Software Architecture, Technical Leadership, OpenAI, Analytics, Prompt Optimization, Written Communication, Interviewing, OpenAI Agents SDK, LoRa, Reinforcement Learning, AI Project Management, Fine-tuning, Demand Forecasting, LLM Fine-tuning, IT Strategy, Monitoring, Large Language Model Operations (LLMOps), Hugging Face, Workflows, Model Deployment, Time Series Forecasting, Bayesian Statistics, FastAPI, LLM Reasoning, Neural Networks, Scalable Vector Databases, Data Visualization, Cloud, A/B Testing, Data Annotation, Attention to Detail, Code Review, Screeners, Team Leadership, Team Management, SFTP, GRPO, SHAP, NER, DBSCAN, MLflow, CI/CD Pipelines, Research, Physics, Applied Mathematics, Linear Algebra, Supervised Fine-tuning (SFT), System Design, Market Segmentation, Churn Analysis, Sentiment Analysis, Natural Language Processing (NLP), Deep Learning, Deep Neural Networks (DNNs), Biotechnology, Clustering, Presentations, Embeddings, Diffusion Models, Generative Models, Mutual Funds, Portfolio Analysis, Portfolio Management, Investments, AI System Design, Claude Code Configuration, Context Management, Generative AI Architecture, Production Reliability, Structured Output, AI Tools, Tool & MCP Design, Semantic Caching
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring