// Languages
function listLanguages()
{
let languages = [
"Python",
"JavaScript",
"R",
"SQL",
"Scala",
"C++"
];
console.log(languages);
}
Building AI Systems That Matter
I Am An
AI Engineer Building Solutions That Matter
Founding AI Engineer building LLM-powered systems at Avenio. Previously shipped AI solutions at World Bank. I specialize in RAG pipelines, multi-agent systems, and taking AI from prototype to production.
What I've Built
- RAG pipeline for 10K biomedical documents with 40% faster retrieval at Avenio
- Multilingual AI assistant processing 478 legal documents at World Bank (90% accuracy, 3.2s response)
- Multi-agent systems handling 100K queries with persona-driven orchestration
- Production deployments on AWS, GCP with Docker/Kubernetes achieving 99.9% uptime
MS in Data Science from George Washington University (GPA: 3.97)
Expert in NLP, LangChain, vector databases, and cloud deployment (AWS/GCP/Azure)
Hands-on with PyTorch, FastAPI, Docker, and building AI systems that deliver real value
6
Languages
8
Tools
1
Certificates
12
Projects
Skills I've Developed and Mastered
I'm hands-on with everything from machine learning and cloud infrastructure to user experience and product thinking. I specialize in building AI systems powered by large language models and retrieval-augmented generation, taking projects from prototype to production deployment across AWS, GCP, and Azure.
- Large Language Models & Agentic AI Systems
- Cloud Deployment & MLOps
- Machine Learning & Deep Learning
- Retrieval-Augmented Generation & Context Engineering
Work Experience
Apr 2025 – Present
Founding AI Engineer, Avenio Corporation, San Francisco, CA
• Architected RAG pipeline for 10K biomedical documents, improving retrieval speed 40% using LangChain, Qdrant, and PGVector.
• Built persona-driven AI agents processing 100K queries at 4s latency via multi-agent orchestration.
• Implemented prompt versioning and evaluation hooks, reducing hallucinations and improving reliability.
• Delivered production pipelines with observability dashboards and AWS deployments via Docker/Kubernetes, ensuring 99.9% uptime.
Jan 2025 – May 2025
AI Engineer (Capstone Fellow), World Bank Group, Washington, D.C.
• Improved access to legal information with a multilingual AI assistant, achieving 90% accuracy.
• Processed 478 legal documents using Python (scraping, OCR, and cleaning) to create a structured, searchable dataset.
• Enabled local language responses to make legal content more accessible to non-English speakers.
• Deployed the assistant on GCP using Docker, ensuring scalable access and average response time of 3.2 seconds.
Aug 2024 – May 2025
Assistant Data Scientist, The George Washington University, Washington, DC
- Built end-to-end ML pipelines for classification, NLP, and retrieval projects using Python, PyTorch, and scikit-learn, conducting model evaluation (precision, recall), hyperparameter tuning, and performance analysis.
- Mentored 13 student teams on AI projects, supporting research on LLM fine-tuning, prompt engineering, and NLP frameworks, conducting code reviews to ensure quality.
Dec 2022 – Jun 2023
Data Science Engineer Intern, Amrita Vishwa Vidyapeetham, India
- Performed data preprocessing and manipulation using Python, Pandas, and NumPy for large-scale datasets in funded research projects.
- Built ETL pipelines and created interactive dashboards using Tableau, Power BI, and SQL for business insights and stakeholder reporting.
- Conducted exploratory data analysis and visualization to support data-driven decision-making across departments.
Dec 2021 – Jun 2023
Data Science Engineer, Cognizant Technology Solutions, India
- Built Fraud Claim Detection system using scikit-learn and PyTorch, improving accuracy through feature engineering and anomaly detection.
- Developed Customer Complaint Routing system using NLP and text classification to automatically categorize and route customer issues, reducing manual routing time by 60%.
- Designed and deployed ML and ETL pipelines using MLflow, Docker, and AWS (EC2, SageMaker, Glue), with data analysis and visualization using SQL, Python, and Tableau.
Education
May 2025
Master of Science, Data Science, The George Washington University, Washington DC.
Relevant Coursework: Data Mining, Data Warehousing, Visualization of Complex Data, Machine Learning, Cloud Computing, Algorithm Design, NLP for Data Science, Linux for DevOps, Foundational Pedagogy GradAsst.
Bachelor of Technology, Computer Science, Amrita Vishwa Vidyapeetham, India.
Relevant Coursework: Natural Language Processing, Machine Learning, Distributed Systems, DBMS, Time Series Analysis, Mining of Massive Datasets, Social Network Analysis.
Projects
Autonomous Research Agent
Built autonomous research agent using LangGraph and tool calling with ReAct pattern, performing multi-iteration web searches and generating cited reports with Tavily API.
BioGraphRAG: Knowledge Graph RAG System
Architected Graph RAG system using Neo4j and LangChain, transforming unstructured biomedical documents into queryable knowledge graphs with 12+ entity types for semantic search.
MediVoice: Real-Time Medical Voice Assistant
Engineered real-time voice AI system for medical consultations using Deepgram STT, Groq Llama 3.3 70B, and ElevenLabs TTS, achieving sub-second latency with WebSocket streaming.
NeXT Mission: AI Assistant for U.S. Veterans
Built full-stack application with Next.js frontend and Django/Python backend integrating LLM APIs for real-time recommendations, optimized for latency and scalability. Runner-up at Meta AI Track.
Finance Advisor Agent
AI-powered financial advisory agent using LangChain and OpenAI with Yahoo Finance API integration for personalized investment insights and real-time market analysis.
AI Code Review SDK
Python SDK for AI-powered code review automation with intelligent suggestions using OpenAI, AST parsing, and CLI tools for seamless developer workflow integration.
Languages & Tools
// Clouds
function listClouds()
{
let clouds = [
"Amazon Web Services (AWS)",
"Google Cloud Platform (GCP)",
"Microsoft Azure"
];
console.log(clouds);
}
// NLP and LLMs
function listNLPandLLMs()
{
let nlpAndLLMs = [
"LangChain",
"LlamaIndex",
"Hugging Face",
"OpenAI GPT",
"Transformers",
"BERT",
"spaCy"
];
console.log(nlpAndLLMs);
}
// ML Frameworks & Tools
function listMLFrameworks()
{
let mlFrameworks = [
"TensorFlow",
"PyTorch",
"Scikit-learn",
"Keras",
"XGBoost",
"Pandas",
"NumPy"
];
console.log(mlFrameworks);
}
// Frameworks
function listFrameworks()
{
let frameworks = [
"Docker",
"Kubernetes",
"FastAPI",
"Django",
"React",
"Next.js",
"Streamlit",
"Spark",
"Kafka"
];
console.log(frameworks);
}
// Vector Databases & DevOps
function listVectorDBs()
{
let vectorDBs = [
"Qdrant",
"Pinecone",
"FAISS",
"Weaviate",
"PGVector",
"GitHub Actions",
"PostgreSQL"
];
console.log(vectorDBs);
}
BioGraphRAG: Biomedical Knowledge Graph System
Abstract: Graph RAG system transforming biomedical documents into queryable knowledge bases with 12+ entity types.
Tools & Technologies Used: Groq Llama 3.3, Neo4j, LangChain, Streamlit, Docker.
Conclusion: Production-ready system for natural language queries over medical knowledge graphs.
SearchRank AI: Neural Search & Ranking System
Abstract: Intelligent search system with neural ranking using BERT for semantic search and learning-to-rank algorithms.
Tools & Technologies Used: PyTorch, BERT, FAISS, Learning-to-Rank.
Conclusion: Production-ready neural search engine for personalized retrieval and ranking.
TripFlow AI: Multi-Agent Travel Planner
Abstract: Built a custom AI travel planner using OpenAI Agent Kit with modular workflows combining classification and itinerary agents.
Tools & Technologies Used: OpenAI Agent Kit, Flight Widget, MCP Servers, Visual Workflow Editor.
Conclusion: Created a real-time flight search and travel planning system with drag-and-drop UI for rapid prototyping and seamless agent orchestration.
JobWatch AI: Automated Job Application Tracker
Abstract: Built an AI-powered automation that tracks job applications while I sleep, processing emails at 1 AM daily using OpenAI to extract job details.
Tools & Technologies Used: Make.com, OpenAI GPT, Gmail API, Google Sheets, Slack, No-Code Automation.
Conclusion: Automated job tracking workflow eliminates duplicates, provides Slack alerts, and maintains a fully updated application tracker with zero manual effort.
NeXT Mission: AI-Powered Veteran Support
Abstract: An AI-powered platform supporting military veterans transitioning to civilian life with a resume builder, chatbot, mentor discovery.
Tools & Technologies Used: Python, Django, Vue.js, Meta LLaMA 4, Groq API, MongoDB, SerpAPI.
Conclusion: A unified platform showcasing Meta's models for real-world impact, developed for the ai+expo hackathon.
Multi-Agent Financial Analytics Engine
Abstract: Built a multi-agent AI bot for real-time financial insights using OpenAI, Groq, and Yahoo Finance APIs.
Tools & Technologies Used: Python, FastAPI, OpenAI API, Groq API, Yahoo Finance API.
Conclusion: Delivered a web-based solution that automates financial analysis and insight generation.
AI Meets Law: Transforming Legal Research with RAG and Fine-Tuned LLM
Abstract: Integrated RAG and fine-tuned LLMs for efficient and precise legal research.
Tools & Technologies Used: Python, Streamlit, FAISS, GPT fine-tuning, SentenceTransformer
Conclusion: Delivered an AI solution for precise legal research with user-friendly interaction.
Text Classification Using SciBERT
Abstract: Multi-label text classification with SciBERT and a multi-head architecture for scientific abstracts.
Tools & Technologies Used: Python, SciBERT, PyTorch, LSTM, CNN
Conclusion: Delivered a high-performing text classifier leveraging SciBERT and innovative deep learning techniques.
Data-Driven Insights for Apple Stock Market Predictions
Abstract: Developed time-series forecasting models to analyze Apple Inc.'s stock data spanning nearly four decades.
Tools & Algorithms Used: Python, LSTM, XGBoost, Voting Regressor, ARIMA
Conclusion: Improved accuracy in stock predictions through feature engineering, hyperparameter optimization, and ensemble techniques.
Advanced Data Visualization of Los Angeles Crime Patterns
Abstract: Created a data visualization platform for analyzing Los Angeles crime trends using a 900K record dataset.
Tools Used: Tableau, Python
Conclusion: Enhanced law enforcement strategies through interactive dashboards highlighting key geographic and temporal patterns.
Credit Card Churn analysis
Abstract: Developed a churn prediction model to enhance customer retention in the credit card industry.
Tools & Algorithms Used: R, Random Forest, Logistic Regression, Decision Tree
Conclusion: Significantly improved customer loyalty strategies through analysis of transaction data.
Supply Chain Management Database Efficiency Analysis
Abstract: Investigated the efficiency of various databases in supply chain management using DataCo's dataset.
Tools Used: MySQL, MongoDB, Neo4j
Conclusion: Optimized database selection for supply chain applications based on data retrieval speed and network pathfinding performance.
Retail Analytics: Utilizing ML for Subscription Insights
Abstract: Predictive analytics framework using Random Forest and Neural Network models for Customer Shopping Preference analysis.
Tools & Algorithms Used: Python, Random Forest, Neural Network
Conclusion: Achieved 97% prediction accuracy in customer subscription behavior through advanced data normalization, categorical encoding, and hyperparameter optimization techniques.
Indoor Multi-Camera Human Tracking
Abstract: Developed a state-of-the-art multi-camera people detection and tracking system for indoor environments.
Tools Used: Python, OpenCV, TensorFlow, Keras
Conclusion: Achieved substantial performance improvements in open environments by successfully addressing occlusion challenges.
Achievement Hub
AWS Certified Cloud Practitioner
Earned the AWS Certified Cloud Practitioner certification, demonstrating foundational knowledge of AWS Cloud services.
Validated expertise in cloud fluency, AWS services, and ability to identify essential AWS tools for cloud-focused projects.
Founding Engineer – Avenio Corporation
Joined Avenio Corporation as a Founding AI/ML Engineer, building real-time agentic AI assistants for biomedical research.
Developed scalable AI pipelines using LLMs, LangChain, and AWS to deliver fast, accurate insights from scientific literature.
AI+ Expo Hackathon (Meta Track – Runner-Up)
Secured 2nd place in the Meta Track at the AI+ Expo Hackathon for building NeXT Mission, a veteran-focused AI assistant powered by Meta's LLaMA 4.
Recognized for impactful use of generative AI and full-stack deployment under tight hackathon timelines.
AWS Summit & Expo – Washington, DC
Participated in the AWS Summit, engaging with cloud professionals and hands-on workshops on GenAI, ETL pipelines, and cloud-native tools.
Explored innovations in AI/ML infrastructure and broadened my knowledge in scalable cloud deployment.
Graduation – Master's in Data Science (GWU)
Graduated with a Master of Science in Data Science from The George Washington University with a 3.97 GPA.
Honed skills in machine learning, NLP, and statistical modeling, laying the foundation for applied AI research.
World Bank – AI Project Presentation
Presented a GenAI legal assistant project at the World Bank, applying RAG pipelines and LLMs to improve access to legal resources.
Earned recognition from mentors and peers for innovation, clarity, and potential policy impact.
Graduate Instructional Assistant – Data Science Capstone
Guided undergraduates through their end-to-end capstone projects—from idea generation to final presentations. Grateful to Professor Sushovan Majhi for his mentorship and trust throughout the semester.
Fine-Tune Your LLM
Through this course, I honed advanced skills in fine-tuning large language models to suit specific business and technical needs.
RAG and Fine-Tuning Explained
This course enhanced my understanding of Retrieval-Augmented Generation (RAG) and fine-tuning techniques for LLMs, showcasing their practical applications.
TensorFlow: Working with NLP
In this course, I acquired hands-on skills in TensorFlow for building natural language processing (NLP) models efficiently.
GPT-4 Foundations: Building AI-Powered Apps
This course provided me with foundational knowledge to build AI-powered applications using GPT-4, focusing on software development and generative AI principles.
Introduction to Generative AI with GPT
This achievement highlights my foundational knowledge in the principles of generative AI using GPT models, focusing on their capabilities and application scenarios.
Generative AI: Working with Large Language Models
I gained advanced skills in large language models (LLM), natural language processing (NLP), and generative AI.
Introduction to Prompt Engineering for Generative AI
I gained essential skills in natural language processing, generative AI & Prompt Engineering.
Generative AI: Introduction to Large Language Models
I gained valuable insights into the architecture and applications of large language models and generative AI.
Introduction to Large Language Models
I gained foundational knowledge in the architecture and functionality of large language models (LLMs).
Learning Design Thinking
I acquired advanced competencies in design thinking methodologies for the effective visualization and interpretation of complex datasets.
Design Thinking: Prototyping
I acquired advanced competencies in design thinking methodologies and prototyping techniques.
Student Athletic Assistant at George Washington University
This picture reflects my role as a Student Athletic Assistant at George Washington University, where I ensured excellent guest services and resolved attendee concerns at basketball events.
DataCamp Intermediate SQL
This certification from DataCamp affirms my intermediate skills in SQL, focusing on complex queries and database manipulation for data analysis.
GitHub Stats & Activity
Featured Repositories
Loading repositories...
Contact Me