Anmol Kumari
Data Engineer & AI Builder. 5+ years turning complex problems into scalable systems - from ML pipelines to production-grade applications.
Specialized in LLMs, data infrastructure, and backend systems. Currently building AI solutions at scale.
Education
Master of Science in Software Engineering
Carnegie Mellon University
Coursework
Bachelor of Technology in Computer Science
Guru Gobind Singh Indraprastha University
Coursework
Skills & Expertise
Additional expertise: Redis, Elasticsearch, Kafka, AWS, GCP, Azure, gRPC, Protocol Buffers, and more.
Work Experience
Research Assistant
ResearchCarnegie Mellon University
- Tmind AIBuilding Tmind AI and Agentic Framework for Mental Health: Full-stack development of AI therapist simulation platform with LLM fine-tuning, RAG-powered context retention, hallucination mitigation, and Google Cloud infrastructure for therapist training - live at tmind.ai
- Distributed SystemsDistributed Systems: Built Instagram-scale social platform demonstrating distributed system architecture principles - implementing scalable consensus algorithms (Raft, 2PC), microservices with Kubernetes orchestration, and fault-tolerant patterns using Saga-pattern coordination for real-time feed delivery
- ML Teaching AssistantMachine Learning Teaching Assistant: Mentoring graduate students in bridging ML theory and adversarial security, covering mathematical foundations of model resilience, robust optimization for deep neural networks, and generative security through diffusion models - facilitating labs on threat modeling and defensive monitoring to build resilient AI systems
Applied AI Engineer
7-Eleven
- Architected GenAI-powered executive dashboard using LangChain SQL agents to enable natural language querying over 7-Eleven's complex data schema - converting business questions directly into optimized SQL queries
- Implemented comprehensive RAG systems with advanced schema management and retrieval optimization, reducing hallucinations through context engineering and prompt optimization for 95%+ query accuracy
Data Engineer
7-Eleven
- Engineered end-to-end production deployment on Kubernetes with real-time monitoring, scaling, and API optimization for sub-second query response times handling thousands of concurrent requests
- Designed and developed evidence generation engine for chargebacks and disputes, processing 1 billion+ records and improving win rate from 5% to 45%, saving $1.5 million annually
- Built multi-source data pipelines in Databricks integrating AMEX, Worldpay, HPS, Discover, Mastercard, and Visa card networks with PySpark optimization and Azure cloud infrastructure
- Awarded Star of the Quarter, Special Contribution Award, and multiple Spot Awards for technical excellence and business impact
Data Engineer
Nationwide
- Built end-to-end PySpark pipelines to flatten, transform, and load XML data into Snowflake for Tableau reporting dashboards
- Optimized Apache Spark jobs with fine-tuned CPU and memory configurations, improving insurance data retrieval speeds for policy and claims histories
- Designed data harmonization workflows stored in Data Lake and Databricks for real-time processing in high-volume environments
Data Intelligence Intern
American Express
- Developed ensemble machine learning model integrating XGBoost and Random Forest, achieving 70% increase in recommendation accuracy
- Implemented SMOTE technique to address class imbalance in offer redemption data for improved model performance
- Identified key customer behavior metrics to fine-tune targeting strategies for personalized marketing campaigns
Projects
Job360 - Multi-Agent Resume Platform
Multi-agent AI system for intelligent resume customization with Central Orchestrator coordinating specialized agents: Profile Parser for resume extraction, Job Understanding for description analysis, Rewrite & Tailor for content optimization, and Document Assembly for final generation. Features job search integration, LinkedIn referral message generation, and interview prep guides - built with FastAPI backend and React frontend.
Method - Retail AI Solution
Hackathon-winning AI-powered retail solution for luxury consumer segment. Implemented intelligent inventory management, personalized shopping recommendations, and customer behavior analytics to optimize retail operations.
Tmind.AI
Distributed system simulating realistic psychotherapy sessions with advanced LLMs for therapist training. Full-stack development spanning LLM fine-tuning, RAG systems with hallucination control, and Google Cloud infrastructure. Designing experimental frameworks to evaluate AI's impact on therapeutic education and clinical outcomes.
Chat with Anmol Agent
Personalized AI conversational agent trained on my professional profile and expertise. Handles queries about my background, projects, and skills using retrieval-augmented generation with context-aware responses.
Multi-Agent Deep Research Systems
Multi-agent AI research system orchestrating clarification, planning, search, and writing agents with quality assurance loops. Implements parallel search optimization and structured Pydantic models for cross-agent communication with mandatory quality scoring feedback.
Customer Segmentation & Behavioral Analysis
End-to-end ML pipeline for e-commerce customer segmentation using K-Means clustering on 5,630 customers across 20 behavioral features. Engineered production-ready segments driving personalized homepage layouts, recommendation engines, promotion targeting, and notification scheduling - enabling real-time personalization with lightweight model scoring in milliseconds.
Accelerometer-based Workout Detection
Machine learning system that analyzes accelerometer sensor data to automatically detect and classify different types of workouts. Achieves real-time activity recognition with high accuracy using signal processing and neural networks.
Tech News LLM Application
Automated tech news aggregation and summarization platform leveraging large language models to curate, categorize, and generate personalized summaries of latest technology news and trends. Features AI-powered filtering for domain-specific news (AI, Cloud, Startups, etc.), personalized reading style adaptation, and intelligent categorization to deliver relevant content tailored to individual preferences and expertise areas.
Achievements & Certifications
YC Agentic AI Hackathon Winner
YC Hackathon Presentation
IEEE Evolution 2025 - Speaker Editor
MLOps Summit
Master of Ceremonies - 6th GenAI Summit 2025
Special Contribution Award
7New Diversity Core Team
Panelist
TechGig Geek Goddess
Fraud Analytics Workshop
Method's Hackathon
Bright Idea Award
Star Award Recognition
Rank 21 - 2018
Top 10 India - KLYES
Hackathon Speaker
College President
Recognitions & Awards
Winner, Y Combinator Agentic AI Hackathon - CMU
Won the Y Combinator Agentic AI Hackathon at Carnegie Mellon University by building a high-impact intelligent agent solution.
Gen AI Summit - Master of Ceremonies
Master of Ceremonies for 6th GenAI Summit 2025
Best Idea Award - CMU
Carnegie Mellon University - Best Idea Award 2025
Best Engineered Product Award
Method - Recognition for engineering excellence in product development
Method Hackathon - Third Place Winner
Third place winner in Method's product hackathon
IEEE Evolution Conference - Speaker Editor
Speaker Editor for IEEE Evolution Conference
Star of the Quarter Award - 7-Eleven
Recognition for outstanding contributions and exceptional performance
Special Contribution Award - 7-Eleven
Multiple Spot Awards for data engineering excellence and technical innovation
7New Core Team - Executive Women Network
Member of 7New - A network of executive women fostering diversity and inclusion initiatives
TerraBot - Springer Publication
Multi-tasking AI published in Springer - Advanced Intelligent Systems. Demonstrates multitasking capabilities across natural language processing, computer vision, and decision-making systems.
AIR 21 & GEEK GODDESS 2018
Ranked among top 1,05,000 contestants in India's competitive tech competition
All India KYES - Exchange Program
Selected for prestigious All India KYES student exchange program