Tanmay VaijFull Stack & Systems
Software Portfolio

Engineering Case Studies & Systems

A showcase of distributed orchestration engines, edge telemetry collectors, caching proxies, and open-source tooling built with high performance and reliability.

SchemaForge: Zero-Downtime DB Migration Engine
Open Source
50k+ Docker Pulls · CI Integration

SchemaForge: Zero-Downtime DB Migration Engine

SchemaForge statically analyzes raw SQL migration scripts and ORM generated migrations to detect dangerous DDL operations like non-concurrent index creation or table-rewriting default values.

Key Technical Architectural Features:
•15+ built-in lint checks for PostgreSQL locking hazards
•Integrates as a native GitHub Action with PR bot reports
•Supports PostgreSQL 12 through 17
Go
PostgreSQL
GitHub Actions
CLI
Web App
450ms → 12ms p95 Latency Reduction

TurboStash: Intelligent Edge Cache Middleware

TurboStash sits between client CDN edges and origin database clusters, managing cache stampede prevention, atomic tag invalidations, and distributed lock leases.

Key Technical Architectural Features:
•Single-flight request deduplication preventing cache stampedes
•Automated cryptographic signature verification on cache flushes
•Zero-config integration with Next.js revalidateTag API
TypeScript
Redis
Cloudflare Workers
Next.js
DevOps & Cloud
40M+ Spans/Day Processed · 99.99% Uptime

CloudPulse: Zero-Overhead Edge Telemetry

CloudPulse aggregates synthetic monitoring, Core Web Vitals, and serverless runtime latency across 30+ edge regions. Featuring instant drill-downs, trace waterfall visualizations, and anomaly detection.

Key Technical Architectural Features:
•ClickHouse columnar storage engine with automated retention policies
•Interactive flamegraph trace analyzer with DOM correlation
•Multi-tenant RBAC with granular API token scopes
Go
Next.js
ClickHouse
Tailwind CSS
eBPF
AI & Systems
1,200+ Stars · 250+ Production Deployments

AgentFlow: Distributed Autonomous Agent Framework

AgentFlow allows engineering teams to construct resilient, self-healing LLM workflows. Built on a Directed Acyclic Graph (DAG) state engine, it supports tool calling, parallel worker execution, human-in-the-loop approval checkpoints, and memory persistence.

Key Technical Architectural Features:
•Sandboxed Python & Bash tool execution engine with memory limits
•Visual DAG execution debugger built with SVG and React Flow
•Zero-dependency lightweight client with sub-10ms serialization overhead
TypeScript
Next.js
Node.js
pgvector
Docker

Have a complex systems problem to solve?

I consult on Next.js performance tuning, distributed database caching, and AI agent execution runtimes.

Get in touch