SENIOR SOFTWARE ENGINEER
V Gnana
Siva Sai
Building high-performance distributed systems and AI-powered developer tools at scale.
FEATURED WORK
Projects
Observability Agent
Identified a critical pain point from my team's on-call experience: engineers waste valuable time context-switching between multiple tools during incident debugging. Architected and developed an AI-powered debugging platform from scratch that unifies all data sources.
PAIN POINT
During on-call rotations, engineers context-switch between logs, infrastructure dashboards, and metrics tools — slowing down incident resolution when every second counts.
SOLUTION
Built an AI-powered platform enabling engineers to investigate production issues using natural language queries across logs, infrastructure, and metrics in one unified interface.
IMPACT
Demonstrated to engineering leadership; significantly reduces mean time to resolution (MTTR) by eliminating tool-switching overhead during incidents.
Built a full-stack SaaS platform to solve a common DevOps headache: unexpected SSL certificate expirations causing production outages. Designed and developed the entire system from certificate monitoring engine to alerting infrastructure.
✓ CORE FEATURES
- •Full chain validation - leaf to root certificate verification
- •OCSP and CRL revocation checking
- •Weak crypto detection (SHA1, weak RSA keys)
- •5-minute check intervals with customizable alert thresholds
- •Real-time dashboard with certificate health status
🔄 IN TESTING
- •Multi-channel alerts (Slack, PagerDuty, Teams, Email, Webhooks)
🔮 ROADMAP
- •Kubernetes Native - Auto-discover certificates from cert-manager
- •AWS ACM Integration - Monitor certificates in Amazon Certificate Manager
- •CT Log Monitoring - Get alerted on unauthorized certificates
- •Domain Monitoring - Track domain expirations alongside certificates
CAREER
Experience
Pure Storage
Member of Technical Staff 3 (MTS-3)
Jun 2023 – Present
Bangalore, India
Observability & AI Platform
- •Architected an AI-powered Observability Agent enabling natural language debugging across Elasticsearch logs, Kubernetes infrastructure, and Prometheus metrics using MCP servers
- •Built 3 MCP servers (Go + TypeScript): Elasticsearch (7 tools), Kubernetes (9 tools), Prometheus (metrics queries)
- •Demonstrated to engineering leadership; reduced MTTR by eliminating tool-switching overhead
Infrastructure & Performance
- •Designed high-performance traffic monitoring system in Go processing 100,000+ logs/second with minimal latency
- •Migrated Kubernetes clusters from EKS 1.24 to 1.31, achieving 25% AWS cost reduction
- •Deployed and maintained Kafka, Elasticsearch, and Cassandra clusters ensuring 99.9% availability
Automation & Efficiency
- •Designed approval system in Go tokenizing user queries into AST trees, achieving 30-50ms latency for 1000+ rules
- •Created auto-approval system using Decision Trees reducing quote approval volume by 77%
- •Engineered Snowflake data processing solution achieving 70-75% reduction in query latency
DevOps & CI/CD
- •Built certificate expiration alerting system integrating cert-manager and AWS ACM with PagerDuty
- •Developed CI/CD pipelines using GitHub Actions and ArgoCD for Kubernetes deployments
- •Implemented Istio service mesh for blue-green deployments using feature flags
Velotio Technologies
Senior Software Engineer
Aug 2021 – May 2023
Pune, India
- •Owned end-to-end feature development delivering 20% increase in application performance
- •Built and maintained fault-tolerant AWS cloud infrastructure ensuring high availability
- •Reverse-engineered and modernized two undocumented legacy Spring Boot applications
Tata Consultancy Services
Systems Engineer
Jun 2019 – Jul 2021
Pune, India
- •Led infrastructure migration from on-premises to Serverless Architecture for London Stock Exchange Group, reducing operational costs by 15%
- •Architected cloud-native solution using AWS Transfer Family transitioning SFTP servers to AWS
OPEN SOURCE
Contributions
NPM Packages
mcp-dev-kit
Complete testing & debugging toolkit for MCP servers
v0.2.0PROBLEM
Building MCP servers is painful without proper testing tools. Developers struggle to validate tool responses, debug protocol issues, and ensure their MCP servers work correctly with AI assistants before deployment.
USE CASES
- •Unit testing MCP tool handlers
- •Integration testing MCP server responses
- •Debugging JSON-RPC communication issues
- •CI/CD pipeline integration for MCP servers
WHO IT'S FOR
Developers building MCP servers for Claude, GPT, or other AI assistants who need reliable testing infrastructure
deveyes
MCP server giving AI coding assistants visual context
v1.1.0PROBLEM
AI coding assistants like Claude are blind — they can't see what your app looks like, what errors appear in the browser, or how your UI renders. They work purely from code, missing critical visual context.
USE CASES
- •Let AI see your React/Vue/Angular component renders
- •Share browser console errors with AI for debugging
- •Debug visual regressions with AI assistance
- •Get AI feedback on UI/UX without screenshots
WHO IT'S FOR
Frontend developers who want their AI assistant to understand visual context, not just code
@gnana997/node-jsonrpc
Transport-agnostic JSON-RPC 2.0 client and server
v1.0.0PROBLEM
Most JSON-RPC libraries are tightly coupled to HTTP or WebSocket. When building MCP servers or custom IPC communication, you need a JSON-RPC implementation that works over ANY transport — stdio, Unix sockets, TCP, or custom protocols.
USE CASES
- •Building MCP servers that communicate over stdio
- •Creating language servers (LSP implementations)
- •IPC communication between Node.js and Go/Rust processes
- •Microservice communication without HTTP overhead
WHO IT'S FOR
Developers who need JSON-RPC but aren't using HTTP — especially those building developer tools, language servers, or MCP integrations
node-ipc-jsonrpc
TypeScript JSON-RPC client over IPC for Go servers
v0.2.1PROBLEM
Connecting Node.js/TypeScript applications to Go backend services over IPC is unnecessarily complex. Existing solutions are either HTTP-based (overhead) or require custom protocol handling.
USE CASES
- •Node.js frontend calling Go backend over Unix sockets
- •VS Code extensions communicating with Go language servers
- •Electron apps with Go backend processes
- •High-performance local communication without HTTP
WHO IT'S FOR
Developers building TypeScript applications that need to communicate with Go services locally — common in VS Code extensions and CLI tools
node-stdio-jsonrpc
JSON-RPC 2.0 client over stdio (child process)
v0.1.0PROBLEM
Spawning a child process and communicating via stdin/stdout with JSON-RPC is boilerplate-heavy. Managing the process lifecycle, buffering stdio, and handling JSON-RPC framing is tedious.
USE CASES
- •MCP client implementations spawning MCP servers
- •Language server protocol (LSP) clients
- •CLI tools that wrap other CLI tools
- •Building VS Code extensions that spawn external tools
WHO IT'S FOR
Developers who need to spawn a process and talk to it via JSON-RPC over stdio — the standard pattern for MCP and LSP
VS Code Extensions
CassandraLens
Comprehensive Cassandra GUI extension for VS Code
BetaPROBLEM
Working with Cassandra is painful. You either use DataStax Studio (heavy, separate app), cqlsh (no GUI), or expensive third-party tools. There's no lightweight, integrated Cassandra client that lives where developers already work — VS Code.
FEATURES
- •Connection management for multiple clusters
- •Schema browsing (keyspaces, tables, columns, indexes)
- •CQL query editor with syntax highlighting
- •Query result visualization
- •Kubernetes cluster integration (connect to Cassandra in K8s)
USE CASES
- •Explore Cassandra schemas without leaving VS Code
- •Run quick CQL queries during development
- •Connect to Cassandra running in Kubernetes
- •Debug data issues alongside your code
WHO IT'S FOR
Backend developers working with Cassandra who want a lightweight, integrated database client instead of switching to external tools
Ollama Copilot
Local LLM-powered code completion and chat
ReleasedPROBLEM
GitHub Copilot and similar tools send your code to the cloud. For developers working on proprietary code, air-gapped environments, or those who simply prefer privacy, there's no easy way to get AI code assistance locally.
FEATURES
- •AI-powered code completion using local LLMs
- •Interactive chat with context from your codebase
- •Works with any Ollama-supported model (CodeLlama, Mistral, DeepSeek)
- •Zero data leaves your machine
- •No subscription required
USE CASES
- •Code completion on proprietary/sensitive codebases
- •AI assistance in air-gapped environments
- •Experimenting with different local LLMs for coding
- •Privacy-conscious development
WHO IT'S FOR
Developers who want AI code assistance but can't or won't send code to the cloud — security-conscious teams, enterprise developers, or anyone who prefers local-first tools
🚀 COMING SOON: Evolving into Archon
A VS Code extension with full codebase semantic understanding for local-first AI coding assistance.
ARCHITECTURE
- •Tree-sitter for multi-language parsing
- •SQLite for symbol storage
- •ONNX / local embedding models (Ollama, vLLM)
- •Local LLMs via Ollama, vLLM, llama.cpp
Currently testing on Kubernetes codebase with Qwen-3 model. Working through tool call issues with local models. Updates coming soon.
OTHER WORK
Projects
Telegram Vault Bot
DevOps · Go
Simplifies HashiCorp Vault management via Telegram bot, automating unseal, rekey, and key distribution with real-time updates.
↳ Secure key distribution among authorized users
Decentralized Cache
Distributed Systems · Go
Distributed cache system using Raft consensus algorithm ensuring scalability, consistency, and fault tolerance.
↳ Production-ready distributed caching
Load Balancer
Infrastructure · Go
Configurable load balancer supporting Weighted Round Robin and Round Robin strategies with health checks.
↳ Easy configuration via YAML file
1 Billion Rows CLI
Performance · Go
CLI tool processing 1 billion rows in ~10 seconds to calculate temperature statistics using optimized algorithms.
↳ Demonstrates high-performance data processing
TECHNICAL
Skills
Languages & Runtime
Infrastructure & DevOps
Databases & Messaging
Frameworks & Tools
CREDENTIALS