Engineered a highly concurrent, RESP compliant Redis-like in-memory key-value store in Go, leveraging optimized mutex-based synchronization to ensure strong consistency and thread-safe transactions across high-throughput read/write workloads.
Implemented a leader–follower replication architecture, enabling asynchronous data propagation, improved readscalability, and fault-tolerant recovery, while preserving data durability under concurrent access patterns.
Achieved 64.4K GET ops/sec and 60.3K SET ops/sec with sub-millisecond average latency (0.46ms GET, 0.62msSET) and p99 latency under 1.8ms, demonstrating efficient memory management and stable tail latency under load.
PROJECT FILE
Doclytics - AI Contract Analytics
CODE NAME
DOCLYTICS
DESCRIPTION
Developed an AI-driven analytics platform to extract clauses from DocuSign contracts using automated pipelines with custom Fivetran connectors and BigQuery.
Orchestrated multi-agent LLM workflows using Google ADK to parse unstructured legal documents, securely exposing the generated insights and analytics via RESTful endpoints.
Constructed an interactive analytics dashboard with dynamic client-side rendering, enabling users to compile and visualize complex data relationships instantly through natural language querying.
PROJECT FILE
CodeTracker
CODE NAME
CODETRACKER
DESCRIPTION
Built a full-stack platform to help programmers monitor and analyze their daily coding activity on Codeforces.
Utilized Prisma ORM and PostgreSQL to handle user data, coding logs, and preferences efficiently.
Added features such as favoriting problems, creating custom tags, and filtering solved problems to enhance user productivity.
PROJECT FILE
Real-Time Chat Application
CODE NAME
NEXUS CHAT
DESCRIPTION
Engineered a low-latency messaging system using WebSockets for real-time communication.
Implemented persistent bidirectional connections between client and server.
Designed scalable room-based architecture for isolated communication channels.
Handled concurrent users and secure session management efficiently.