name : Ansar Afsar
role : AI Systems Engineer & Junior AI Developer @ Webdura Technologies
location : India 🇮🇳
focus : AI Agents · RAG · Workflow Automation · Production AI Infrastructure
philosophy : Benchmarks are useful. Operational reliability matters more.
status : 🟢 Building AI that survives productionI work at the seam between AI systems, workflow automation, product thinking, and infrastructure — building systems that don't just demo well but run reliably at operational scale.
My systems handle:
🌍 Multilingual inputs | ⚡ Latency constraints | 🚢 Deployment realities | 🔄 Workflow integration | 📊 Operational scale
 🤖 Orchestration & Multi-Agent Systems |
 🔎 Retrieval-Augmented Generation |
 ⚡ Workflow Automation Pipelines |
 🛠️ Cloud Deployment & Infra |
🧠 Junior AI Developer — Webdura Technologies | 2026 → Present
Building AI-backed marketing products and AI agent assistants for traditional businesses.
- ⚙️ Designing AI-assisted operational workflows from scratch
- 🧩 Working closely across product ideation and real business pain points
- 📈 Building scalable automation systems around AI primitives
- 🚀 Rapid R&D on deployable AI workflows
🤖 AI/ML Developer — Teamup Consultants | 2025
Built AI workflows and cloud-native systems for recruitment ops across Gulf & Middle East markets.
- 🧠 Generative AI workflows for hiring pipelines
- 🔐 Authentication-integrated AI systems
- ⚡ Rapid AI prototyping infrastructure
🛡️ AI Module Lead — Tienext Corporation | 2025
Owned NLP moderation infrastructure deployed at production scale on AWS.
- 🌍 Multilingual hate-speech & toxicity detection
- ⚡ Real-time moderation pipelines
- 🐳 Dockerized self-hosted inference systems
- ☁️ Production-scale deployment workflows
👁️ AI Developer — Zeex AI | 2025
Built computer vision systems across surveillance and analytical domains.
- 🚨 Theft detection · 🚦 Traffic analysis · 🛰️ Satellite imagery processing
- 🏗️ Few-shot learning pipelines using Vision Transformer backbones
📊 Team Manager, Content/Data — Bookdio | 2024 → 2025
Led AI-assisted content optimization and analytics operations.
- 📈 Scaled organic impressions: 2.43K → 477K 🚀
- 🧠 Built analytics-driven operational processes
- 🤝 Managed AI-assisted content systems at scale
Current technical bets and directions I'm going deep on:
| Area | What I'm Exploring |
|---|---|
| 🤖 Agent Orchestration | Multi-agent systems, tool use, memory layers, self-correcting pipelines |
| 🔎 RAG & Retrieval | Hybrid search, re-ranking, structured + unstructured data retrieval |
| ⚡ AI Workflow Automation | n8n, LangGraph, event-driven AI pipelines for business ops |
| 🏢 Business-Facing AI | AI systems for industries that still run on manual processes |
| ☁️ Production Infra | Dockerized inference, self-hosted LLMs, latency optimization |
┌─────────────────────────────────────────────────────────────────┐
│ │
│ AI systems that survive production. │
│ AI that fits how businesses actually operate. │
│ AI that automates real workflows, not toy demos. │
│ Infrastructure that creates leverage at scale. │
│ │
│ Benchmarks are useful. │
│ Operational reliability matters more. │
│ │
└─────────────────────────────────────────────────────────────────┘
🚴 Cycling clears my head better than debugging | 🎮 Strategy games are my favorite way to think through systems
📚 Usually reading AI papers, infra blogs, or startup/operator essays | ☕ Most ideas start from overthinking workflows that could be automated
