→
Master of Science
Computer Science and Engineering
Courses
Algorithms
Machine Learning
Security
Computational Linguistics
Data Intensive Computing
Research Assistant: DRONES Lab: Interactive, Agentic AI, RAG, LLM Inferencing for children's language learning.
→
Bachelor of Technology
Computer Science and Engineering
Courses
Data Structures
Operating Systems
DBMS
Networks
Computer Architecture
ML
OOPS
Highly skilled Generative AI Researcher and Software Engineer with a Master's in Computer Science, specializing in developing advanced AI/ML systems and robust backend infrastructure. Proven ability to drive innovation through multimodal LLM applications, optimize complex systems for significant cost savings (up to 83%), and deliver high-impact solutions serving large user bases (150,000+). Seeking to leverage expertise in AI research, scalable backend development, and project management to contribute to cutting-edge technological advancements.
Generative AI Researcher
Buffalo, New York, US
→
Summary
Led the development of an LLM-driven Science of Reading simulator, integrating multimodal cues and robust GenAI backend systems for adaptive teacher interaction.
Highlights
Built an LLM-driven Science of Reading simulator utilizing multimodal cues (speech, visual aids) with VR and Unity to model student responses, actions, and emotions for adaptive teacher interaction.
Developed the full GenAI and backend stack using LangGraph, FastAPI, Docker, ChromaDB, and RAG, incorporating session memory for context-aligned responses with robust safety evaluations.
Integrated ONNX and PyTorch speech-to-text and phoneme models for real-time, accent-aware transcription, enhancing simulator capabilities.
AI Intern
Duarte, California, US
→
Summary
Developed and optimized LLM evaluation metrics and agentic systems, significantly improving model adoption and operational efficiency for HopeLLM.
Highlights
Developed evaluation metrics for hallucinations, accuracy, and completeness using LLM-as-a-Judge and Chain-of-Thought with Langchain and LangGraph, increasing HopeLLM's adoption by 45%.
Optimized multi-stage OpenAI GPT model usage based on custom metrics, cutting operational costs by 83%.
Implemented an agentic system for tool evaluation and comparison generation, saving over 8 hours per deployment.
Software Engineer
Bengaluru, Karnataka, India
→
Summary
Developed and optimized highly available Node.js REST APIs and backend ETL pipelines, enhancing system performance and developer operations.
Highlights
Developed and maintained multiple user-facing Node.js REST APIs serving 150,000+ people via AWS Lambda and API Gateway, incorporating unit tests and AWS SDK mocking.
Built a fault-tolerant JavaScript ETL pipeline in collaboration with cross-functional teams, ingesting and transforming over 30K external records daily into a data lake for analytics and archival.
Optimized an AWS Keyspaces DB schema, reducing data fetch latency by 70%, and achieved 40% reduced concurrent active AWS RDS sessions with SQL connection pooling.
Integrated AWS Personalize to build a demographic-aware news recommendation engine, significantly increasing user engagement.
Enhanced DevOps CI/CD pipeline with AWS CloudFormation and Jenkins, reducing build and deployment time.
Published by
Springer (International Conference on Artificial Intelligence and Data Science)
Summary
Classified different rice crop diseases from images using DICOM image segmentation, performed GradCAM analysis, and reported efficiencies of various deep learning and computer vision classification models.
Published by
International Journal of Engineering, Applied Science and Technology
Summary
Performed Exploratory Data Analysis using Pandas, Numpy, and Matplotlib, extracted features from unstructured medical reports using NLP techniques, and developed a fine-tuned Question-Answer model with HuggingFace's BERT.
Summary
Developed an AI-driven grading platform that evaluates PDFs, code, notebooks, and videos submissions.
Summary
Built a TCP-based chat system in C++ using sockets for concurrent clients.
Summary
Implemented and optimized fully connected Neural Networks and CNNs, including VGG-13 and ResNet-34 in PyTorch for multi-class and binary classification.
Node.js, JavaScript, Python, Pytorch, C++, C, Angular, FastAPI, Flask, Django, Java, JQuery, TypeScript, React, Spring Boot, Tensorflow, Keras, Langchain, LangGraph, LLMs.
MySQL, PostgreSQL, MongoDB (NoSQL), CosmosDB, Redis, Hadoop, Spark.
Lambda, S3, CloudWatch, RDS, DynamoDB, Cassandra Keyspaces, IAM, Step Functions, Secrets Manager, Glue, API Gateway, EC2, CloudFormation, SQS, ElasticSearch, Elasticache, Personalize.
Git, Github, Azure Devops, Jenkins, Jira, Docker, Kubernetes, Kafka.