Building a real-time voice AI tutoring application for Spelling Bee practice for 1000+ active users. Solo engineering the entire product from system design to deployment: Python backend with LiveKit Agents SDK, Next.js frontend, PostgreSQL on Supabase, serverless infrastructure on Vercel and LiveKit Cloud.

• Built real-time voice pipeline (VAD, STT, LLM, TTS), achieving sub-1s end-to-end latency using LiveKit, WebRTC, and AssemblyAI

• Designed function calling tools that query a word database directly, eliminating LLM hallucination during word presentation and spelling evaluation

• Implemented layered architecture with repository pattern, service layer, and pluggable provider system for runtime swap of AI services, tracking detailed metrics for unit economics and personalized user analytics

• Reduced speech-to-text costs by 40-50% through Silero VAD integration, filtering silence before API transcription

#VoiceAI #RealTimeSystems #SoftwareEngineering #AIEngineering #WebRTC #LLM #EdTech #Startup #FullStack #Python

Built production-grade AI applications and data pipelines as part of the Big Data Systems and Intelligent Analytics program. Developed full-stack systems spanning data engineering, AI/ML, and cloud deployment while mentoring 50 graduate students on technical projects.

• Built multi-agent AI systems using LangGraph with RAG pipelines, semantic caching, and vector databases achieving 92% retrieval accuracy

• Developed ELT pipelines processing 279K financial filings using Airflow, dbt, and Snowflake with medallion architecture

• Implemented Snowflake-native CDC using Streams and Tasks for event-driven data processing with 80% compute reduction

• Deployed FastAPI backends and Next.js frontends on GCP and AWS with Docker containerization and GitHub Actions CI/CD

• Designed and delivered 8 technical labs covering data engineering, RAG systems, and agentic AI to graduate students

#DataEngineering #AIEngineering #LangGraph #RAG #Snowflake #Airflow #dbt #FastAPI #Python #FullStack

Application Engineering and Development, INFO5100

Delivered technical support and active mentorship to graduate students in a hybrid curriculum covering Java, object-oriented programming, software engineering principles, and UX design for business application development.

Clarified complex topics in application modeling, software development, and user experience, increasing student assignment scores and overall retention.​

Guided project-based teams in designing scalable, user-focused business apps using Java and OOP methodologies.​

Contributed to grading, classroom management, and constructive feedback, aligning with course goals and enhancing student engagement.​

Collaborated closely with faculty to optimize lesson delivery and refine course content, leveraging personal experience as a former top-performing student.​

Ready to present technical and teaching expertise in software-driven business solutions and academic environments

Built backend services and data pipelines end-to-end for an enterprise B2B supply chain platform serving 150+ clients, and 200+ internal users across procurement, manufacturing, and distribution.

• Developed FastAPI services with Pydantic validation, JWT authentication with RBAC, PostgreSQL transactions, and Redis caching, publishing events to Kafka for downstream processing

• Built Spark batch and Kafka streaming pipelines with broadcast joins, partition pruning, and small file consolidation from 40+ hubs, transforming through medallion architecture into Redshift

• Automated financial reconciliation and supplier payout workflows, reducing manual Excel processing from 8 hours to 30 minutes

• Designed star schemas with SCD Type 2 for historical tracking and wrote optimized SQL for Power BI dashboards serving executive leadership

• Built React components for internal tools and deployed containerized services on AWS with Terraform, Airflow, and GitHub Actions CI/CD

#SoftwareEngineering #FullStack #DataEngineering #FastAPI #React #Python #Spark #Kafka #AWS #PostgreSQL #Airflow