AI automation and agentic-systems builder. I turn brittle AI workflows, generated codebases, and messy integrations into working shipped tools with tests, screenshots, logs, and handoff docs.
Open this week for paid 24-48 hour sprints. Best fit: a concrete workflow, agent, or internal tool that needs to work in production fast. Start from the live sprint menu or email a one-paragraph scope.
- AI workflow rescue: n8n, Make, Zapier, webhooks, retries, CRM handoffs, observability, and failure maps.
- Agentic codebase audit: review AI-generated TypeScript/Python apps, find fragile boundaries, and ship the first cleanup patch.
- Voice-agent reliability: Retell/Vapi/GoHighLevel transcript triage, replay cases, failure tags, and implementation fixes.
- Internal AI tool build: a small LLM-powered tool with live UI, API glue, eval checks, and runbook.
- Live proof homepage
- Voice-agent failure triage tool
- Landing-page leak finder demo
- Agent cost leak checker
- Context OS v2.9.0 GitHub Action
- Private repo audit request
- context-os: CI gate for coding-agent context and cost leaks across Claude Code, Codex, Cursor, and other agentic coding workflows.
- schema-gateway: schema portability and regression guardrails for structured outputs across OpenAI, Gemini, Anthropic, and Ollama.
- glasswall: GitHub-native patch-gap operations for public-fix-to-private-remediation workflows.
- agent-activity-graph: runtime evidence layer for AI agents in enterprise workflows.
Email: sravan272001@gmail.com
