On AI Lounge since Nov 2026

Hamza Agar

@aimtuneMember #5,841DeveloperL2
Software Developer
RAG-powered GenAI chatbots on Azure OpenAI for enterprise customer support
Message 🔒
5
karma
5
projects
0
upvotes
1
followers
1
following
0
masterminds
114
points

Achievements · 3 of 20

Plugged In ×3Connected sources: GitHub, Hugging Face, LinkedIn, X, website.How: Connected 3 sources (GitHub, LinkedIn, X or a website) in SettingsEarned Nov 21, 2026
Plugged In×3Featured
Shipper ×2Projects on your profile.How: Added 5 projects to their profileEarned Nov 21, 2026
Shipper×2
StorytellerPosts published.How: Published 1 postEarned Nov 21, 2026
Storyteller

Pinned projects

chativa
Repo · TypeScript · LitElement · Web Components
▲ 0
A flexible, lightweight chat widget built as a web component with LitElement, so it drops into React, Vue, Angular or plain HTML with a single script tag. Pluggable connectors for DirectLine, WebSocket, SignalR, SSE, HTTP and Mekik (or your own), rich message types (cards, buttons, quick replies, carousels, files, video), inline streaming generative UI (forms, charts, tables), middleware for analytics and slash commands, CSS-variable theming with four window modes, English/Turkish i18n, end-of-conversation surveys and a multi-conversation agent panel for helpdesks. Hexagonal architecture, strict TypeScript, MIT licensed. Live sandbox available.
ChatWeb ComponentsConversational AIGenerative UITypeScript
ClassifAIer
Repo · Python · LangChain · scikit-learn
▲ 0
Text classification with LLM embeddings and classic ML. ClassifAIer turns text into embeddings from any LangChain-supported provider (e.g. OpenAI), then trains a scikit-learn classifier (RandomForest, KNN, etc.) on them to label new text. Simple fit / predict_all / save / load API, so a trained model can be stored and reused. Language-agnostic: the example classifies Turkish news into categories like sports and economy. MIT licensed, pip install ClassifAIer.
Text ClassificationEmbeddingsLangChainPythonOpen Source
converger
Repo · Elixir · Phoenix · PostgreSQL
▲ 0
High-performance, multi-tenant real-time messaging infrastructure built with Elixir/Phoenix. Connects a tenant's channels (webhooks, WhatsApp, WebSocket clients) through conversations: every activity is persisted first, streamed to connected clients in real time, then delivered to routed channels with retries. Pluggable inbound/outbound/duplex channel adapters with fan-out routing, idempotency keys for safe retries, tenant and IP rate limiting, HMAC and WhatsApp/Meta webhook signature verification, a LiveView admin panel, and full observability (OpenTelemetry, Prometheus, Grafana, Jaeger, Loki). k6 load tests plus a chaos test that kills nodes under load: zero lost acknowledged messages in the latest benchmark.
RealtimeMessagingElixirInfrastructureMulti-tenant
ilmek
Repo · TypeScript · .NET · SQLite
▲ 0
An agent graph runtime with durable, journaled human-in-the-loop, for TypeScript and .NET. Side effects run inside steps and are recorded in a journal, so when a flow pauses for a human and resumes, completed actions are replayed from the journal instead of running twice. Interrupts are first-class (multiple, in loops, in parallel), graphs are serializable data, and it supports fan-out, node-directed commands, retries and typed event streaming with token streaming and cancellation. SQLite and Postgres checkpointers. Both implementations pass a shared conformance suite (176 TS + 28 .NET tests).
AgentsHuman-in-the-loopWorkflowTypeScript.NET
mekik
Repo · TypeScript · .NET 9 · WebSocket
▲ 0
Realtime serving layer for ilmek graphs: turns a running agent graph into a live conversation over WebSocket. Clients send user turns and interrupt answers; the server streams back text, generative UI chunks, tool traces and human-approval pauses over one wire protocol. Tool side effects run exactly once across pause/resume, clients replay missed frames after reconnect, and the frontend can expose its own UI capabilities as tools. TypeScript and .NET implementations verified byte-for-byte against shared golden fixtures. Integrates with Microsoft.Extensions.AI, Semantic Kernel, LangChain, Redis and the chativa widget.
AgentsRealtimeGenerative UIWebSocketHuman-in-the-loop

Stack

OllamaClaudeLangfusen8nAmazon BedrockChatGPTLangGraphgpt-ossGitHub CopilotDeepSeekGemmavLLM12 tools

Experience

Software Developer | .NET | Azure | Elixir Enthusiast
Dogus Technology logo
Software Developer Specialist · Dogus Technology
Jul 2022 – PresentCurrent

Recent posts

All posts
Shipped: converger
Agents · 58m ago · ▲ 0 · 0 comments
Shipped: chativa
Agents · 58m ago · ▲ 0 · 0 comments
Shipped: mekik
Agents · 58m ago · ▲ 0 · 0 comments
Shipped: ilmek
Agents · 59m ago · ▲ 0 · 0 comments
Shipped: ClassifAIer
Agents · 59m ago · ▲ 0 · 0 comments