Tejaswini Chavan

- Engineered a hybrid recall engine combining two-stage retrieval (semantic search + BFS graph expansion) with a lifestyle classification system achieving 100% accuracy at 213.9 moments/second, alongside a spaCy + BERT hybrid NER pipeline for entity-aware memory enrichment. - Designed and implemented a multi-signal memory threading engine using semantic similarity (MiniLM-L6-v2), Jaccard entity overlap, and exponential decay temporal proximity, processing 535 moments into 1,024 relationships with 86% thread connectivity on a PostgreSQL-native graph schema. - Built a FastAPI-based autonomous agent service featuring a Priority Agent, Theme Organizer, and Loose End Detector with Supabase-backed persistence, per-user self-evolving priority weights, and vector-based semantic retrieval across a live cognitive memory platform.

- Developed 5+ AI applications and APIs using Next.js, OpenAI, Pinecone, and Stripe API, achieving 98% accuracy in results, as demonstrated by feedback from over 1,000 users. - Developed projects from design to deployment, leading 4+ engineering fellows to use MCP Servers. - Coached by Amazon, Bloomberg, and Capital One engineers on Agile, CI/CD, Git, and microservice patterns.

• Developed applications using Java and Spring Boot, integrating PostgreSQL as the relational database, which enhanced data processing speed by 35% for key reporting features. • Designed and implemented scalable web services with Node.js and Express.js, optimizing MongoDB schemas and leveraging Redis caching, leading to a 200ms improvement in average API response times. • Engineered three specialized Salesforce Lightning interfaces for Sales and Marketing teams, reducing data retrieval time by 35% and accelerating report generation. • Streamlined tax compliance workflows and backend processes by incorporating a rate limiter and utilizing Spring Boot with RESTful APIs, achieving a 70% decrease in tax calculation processing time. • Diagnosed and resolved critical CI/CD pipeline issues by employing Packer, GitHub, and Amazon CodeDeploy, reducing deployment failures by 30% and enhancing team efficiency.

• Partnered with the UX design team to develop an A/B testing framework, increasing user engagement by 20% and improving conversion rates for key product features. • Built a real-time data pipeline with Python and Apache Kafka, enhancing data freshness for stakeholders and maintaining a 99.9% data accuracy rate across dashboards. • Optimized SQL database queries for the client reporting feature, resolving performance bottlenecks and reducing report generation time by 4 seconds per report. • Developed a streamlined error logging system with Sentry, enabling real-time alerts and diagnostics, which accelerated critical issue resolution by 60%.

Courses: 1) INFO 5100 Application Engineer & Development (Java, Java Swing, Db4O) 2) DAMG 6210 Data Management and Database Design (ER Modeling, Normalization, MSSQL, Joins, Triggers, Encryption, MYSQL, MongoDB, Tableau ) 3) INFO 6205 Program Structure and Algorithms in Python (Solving techniques: brute force, recursion, divide-and-conquer, dynamic programming, greedy algorithms, iterative improvement and backtracking) 4) INFO 6150 Web Design & User Experience (HTML, CSS, JavaScript, Bootstrap, MongoDB, Express JS, Node JS, React) 5) CSYE 7280 User Experience Design/Testing UIUX ( UX Research Methods, Prototyping, Wireframes, Figma) 6) CYSE 6200 Concepts of Object Oriented Design (Use of inheritance, composition, interface,Java applets, Low Level & High Level Design) 7) CYSE 6225 Network Structures and Cloud Computing (Cloud architectures, cloud security, cloud governance, cloud virtualization, and cloud capacity, EC2, S3, API Gateway, RDS, VPC Route, Route S3, IAM, Terraform)