Daily workflow
AI coding workflow
Claude Code · OpenAI Codex · Cursor
Planning, implementation, debugging, code review and tests with AI coding assistants. I review outputs against the task and repository checks.
AI experiences & automation · India / remote
Skills & AI toolkit
AI-native workflow. Software engineering underneath.
I use AI assistants to build, investigate and iterate, alongside the code, APIs, tests and deployment work that make an application usable. The labels below distinguish working tools, source examples, experiments and current learning areas.
Working toolkit
General AI assistants
ChatGPT · Claude · Google Gemini
Research, problem framing, technical explanation and prompt iteration. These tools complement source inspection and engineering validation.
Public workflow repositories
Assistant configurations
OpenCode · GitHub Copilot · Google Antigravity · Gemini CLI · Aider
Tool-specific instructions and coding workflow exploration. Repository scope varies by assistant; configuration is not a certification.
Product evidence
LLM application integration
Vercel AI SDK · OpenAI APIs · Anthropic SDK · structured outputs
QuizFlow provides the product example; the Anthropic SDK adapter is an experiment with a documented roadmap.
SDK experiments
Agent orchestration
LangChain · LangGraph · CrewAI · AutoGen · LlamaIndex
Exploration across agent frameworks, prompt composition and graph-style mission workflows. Inspect each repository before treating a capability as complete.
Exploration
Model and provider ecosystem
OpenRouter · DeepSeek · Groq · Ollama · Gemini models
Provider selection, API compatibility and local versus hosted model tradeoffs are part of my AI project interests. No claim of production use of every provider.
Core engineering stack
Full-stack delivery
TypeScript · JavaScript · Python · React · Next.js · Node.js · Express
Application interfaces and APIs around AI features, rather than an isolated prompt or notebook.
Product and broader stack
Data and interface foundations
PostgreSQL · Supabase · Prisma · SQLite · IndexedDB · Tailwind CSS
Local-first browser persistence in QuizFlow; database and responsive UI tools across my wider web stack.
Public source evidence
Testing and deployment
GitHub Actions · Vitest · pytest · Docker · Vercel · Cloudflare Workers
Tests, explicit contracts, deployment checks and handover documentation. Service-level guarantees require an agreed scope.
Project focus
Workflow automation
n8n · Make · webhooks · REST APIs · CRM handoff
Connecting triggers, business rules, AI steps and human review. Engagements begin by examining the current workflow and access requirements.
Developing deeper evidence
Retrieval and reliability
RAG · embeddings · vector search · context engineering · evals · observability
Current direction: grounded retrieval, evaluation datasets, tracing, guardrails, cost controls and clear failure behavior. Not presented as measured production achievements.
How I choose a tool
Start with the task, data boundaries and acceptance criteria. Pick a model and interface that fit the workflow, verify output with source and tests, and keep a usable fallback when an AI service is unavailable.
One conversation, a concrete next step
Let’s build something useful.
Send a role or a short project brief. We can discuss the problem, location eligibility, stack, scope and availability.
Download résumé · Global collaboration · Recruiter & client FAQ