const anish = {
  role: "Full-Stack Developer",
  focus: ["LLMs", "AI Agents", "Automation"],
  stack: ["React", "React Native", "Node.js", "Python"],
  databases: ["MongoDB", "MySQL"],
  currentlyExploring: "System-level engineering",
  openTo: ["Full-time roles", "Freelance", "Early-stage startups"],
  portfolio: "https://anishdharnidhar.dpdns.org/",
};

🔗 Live: prajuririveracres.com

• Enabled direct online reservations and payments for a resort business, reducing dependence on third-party booking channels, with a full-stack platform: JWT auth, Razorpay payments, and a Next.js SSR admin portal for real-time booking management
• Drove inbound leads through an SEO pipeline (sitemap generation, structured metadata) and lead-capture features on a responsive Tailwind CSS frontend

Next.js React Tailwind CSS JWT Razorpay SEO

🔗 Live: sindhuyouthcircle.org

• Enabled a 23-language voice assistant for non-English-speaking NGO community members by designing a from-scratch multilingual language-detection engine (9 Indic scripts, romanized "Hinglish" detection across 6 languages) with zero external NLP libraries
• De-risked model selection by prototyping and benchmarking parallel speech-AI pipelines, cloud LLM vs. fully offline local inference, integrating faster-whisper STT, Gemini / AIRLLM / Ollama, and neural TTS with GPU/CPU auto-detection

Python faster-whisper Gemini Ollama AIRLLM TTS

🎬 Demo: Watch the demo

• Made banking forms accessible to visually impaired users by owning the Flask/Python ML backend with Hugging Face pipelines (flan-t5, NLLB-200, XLM-RoBERTa, MarianMT) for multilingual OCR, translation, and voice-guided form filling
• Restored production stability on Cloud Run by resolving 503/OOM worker-timeout failures (Gunicorn concurrency tuning, Docker config fixes) and authored deployment runbooks that cut team debugging time

Flask Python Hugging Face Docker Gunicorn Google Cloud Run

💡 My AI assistant's server is an old Android phone. Hosting bill: ₹0.

• A personal AI assistant that lives in Telegram and runs on that old Android phone (Termux). Chat in plain language to set reminders, keep notes, store and read files, search the web, edit GitHub repos through pull requests, and read or edit Google Docs, Sheets and Slides
• Built for one owner plus trusted guests, with per-user data isolation and owner-only access to GitHub and Google. LLM tool calling via any OpenAI-compatible API (default Groq gpt-oss-120b), SQLite storage, and a CI/CD pipeline that tests, deploys to the phone over Tailscale + SSH, health-checks, and rolls back on failure

Python Telegram Bot API LLM Tool Calling SQLite Termux Docker GitHub Actions Tailscale

Hiring, building a product, or need a developer who can own the whole thing? I'd like to hear about it.

• 📧 Email: anishdharnidhar07@gmail.com
• 💼 LinkedIn: /in/anishdharnidhar
• 🌐 Portfolio: anishdharnidhar.dpdns.org