class AsadAhmed:
    name        = "Shaik Asad Ahmed"
    role        = ["AI Engineer", "Agent Architect", "LLM Systems Builder", "Backend Developer"]
    location    = "Kurnool, Andhra Pradesh, India 🇮🇳"
    education   = "B.Tech CSE (AI) · GPCET · GPA: 8.5/10 · 2022–2026"
    currently   = "Interning @ HandiCrafts · Building AI Healthcare Chatbot · Built DevCrew (6-Agent Project Builder)"

    def what_i_do(self):
        return [
            "Build multi-agent pipelines that run autonomously",
            "Orchestrate LLMs into real production systems",
            "Deploy AI apps — not just notebooks",
            "Connect NLP → Backend → Automation end-to-end"
        ]

    def philosophy(self):
        return "Build things that solve real problems. Deploy them. Iterate. 🔁"

Describe any project idea in plain English — a 6-agent AI crew analyzes, plans, codes, tests, and delivers a ready-to-run ZIP file. Built in a weekend. Zero cost to run.

▸ 6-agent pipeline: Goal Analyst → Tech Advisor → Planner → Coder → Tester → QA — fully sequential and autonomous ▸ Tech Advisor honestly recommends against its own stack (CrewAI) when simpler tools fit better ▸ Generates complete projects across CrewAI, LangChain, Vanilla Python, Node.js stacks with correct imports ▸ Solved Groq rate limit battles (6,000 TPM) across 6 agents via per-agent crew isolation + timed delays ▸ Delivers ZIP with all code + README + .env.example — zero setup friction for the user

Real-time voice agent for veterinary emergency triage. WebRTC audio streaming + LLM reasoning + neural TTS — deployed and tested across real clinical scenarios.

▸ Sub-2 second end-to-end speech-to-response latency achieved ▸ Handles mid-conversation context switching between emergency & scheduling modes ▸ Validated across 6 clinical scenarios — cats, dogs, birds, street animals ▸ Async tool-calling with real-time slot booking via n8n + Supabase PostgreSQL

Autonomous 5-node agent swarm that classifies, routes, and acts on emails — zero human intervention needed.

▸ 5-node pipeline: Gmail polling → LLM classification → priority routing → calendar detection → Sheets memory ▸ Auto-Unsubscribe: LLM detects spam links, triggers browser automation autonomously ▸ ~70% inbox noise reduction with zero user action ▸ Closed feedback loops for adaptive agent learning over time

Faith meets engineering. An agentic prayer reminder that follows up, checks compliance, and responds with hadith if a prayer is missed.

▸ Reminders sent 15 min before each Salah via Telegram with voice + text ▸ On missed prayer: queries Supabase hadith DB → context-aware motivational response via Claude API ▸ Dual-LLM architecture: Claude API + Groq for A/B response testing ▸ Full-stack onboarding: React.js + Tailwind CSS registration page + Supabase user store

Production-deployed AI ordering system for a live restaurant. Eliminates 10 documented operational failure modes vs paper menus and static QR codes.

▸ LLM-powered dish Q&A + neural voice descriptions (OpenAI TTS nova model) ▸ 4-tier architecture: CDN frontend → PostgreSQL → n8n automation → dual LLM inference ▸ Auto email dispatch to customer + kitchen on every order placement ▸ Live deployed at ~USD 65/month operational cost

Fully autonomous multi-source research curation. Fetches from arXiv, Semantic Scholar, HuggingFace Papers, Google Scholar — ranks, scores, and delivers to Telegram.

▸ 3 interconnected n8n workflows with AI-driven scoring and summarization ▸ Persistent deduplication memory via Supabase PostgreSQL ▸ Real-time Relevant / Not Relevant feedback loop that improves recommendations ▸ Zero manual intervention — weekly AI-generated digest fully automated

Multi-agent AI system for real-time financial transaction analysis, regulatory validation, and automated risk scoring.

▸ NLP + rule-based reasoning for real-time compliance checking ▸ JWT-secured role-based dashboard with REST APIs ▸ CSV/PDF reporting with end-to-end auditability

3-agent swarm that analyzes GitHub activity, resume text, and coding profiles — computes weighted technical readiness scores for placement prediction.

▸ 3 specialized agents: GitHub analyzer + resume parser + coding profile evaluator ▸ FastAPI orchestrator aggregating outputs into unified readiness dashboard ▸ RESTful endpoints for real-time scoring and data delivery

  

🏢 AI Application Developer Intern — HandiCrafts (Sep 2025 – Present)

▸ Building LangChain-powered AI healthcare chatbot with Python + NLP ▸ Improved RAG query resolution accuracy by ~35%, reduced response latency by 40% ▸ Engineered prompt strategies and hallucination mitigation for production reliability

🏛️ Python Developer Intern — IIITDM Kurnool (May 2024 – Aug 2024)

▸ Built voice-based payment system: Python + SpeechRecognition + REST APIs ▸ Implemented QR payments, JWT authentication, real-time transaction orchestration ▸ Agile team environment — Git, debugging, performance optimization