Skip to content
View Abdur-Rafay-AR's full-sized avatar

Highlights

  • Pro

Block or report Abdur-Rafay-AR

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Abdur-Rafay-AR/README.md

Abdur Rafay

AI/ML Engineer building production-grade agentic systems

RAG pipelines, multi-agent systems, and voice AI in production


Background

AI/ML engineer pursuing a Bachelor's in Artificial Intelligence. Previously Member of Technical Staff at Lythe Labs (SF), shipping agent infrastructure and voice pipelines. Shipped 25+ production AI systems spanning RAG pipelines, multi-agent orchestration, and voice agents, backed by six internships across international teams. Open to full-time AI/ML roles.


Stack

Languages Python TypeScript

AI / Agents Groq LangChain RAG

Backend & Real-time FastAPI WebSocket SQLite

Frontend & Infra React Docker AWS


IBM AI Developer Professional Certificate

More on Credly →


Featured work

Project What it does Stack
athar Local-first time-travel debugger and observability engine for multi-agent LLM systems, fork/replay from a mutated step without re-running earlier steps. Published to PyPI (pip install athar) Python, SQLite WAL
relay AST-based async latency auditor for Python agent codebases, finds tool-call chain bottlenecks with no runtime overhead. Published to PyPI as relay-arclat Python
voice-agent-starter-kit Real-time voice agent template built at Lythe Labs: streaming ASR → LLM → TTS pipeline over WebSocket, feels like an actual phone call Python, FastAPI, WebSocket
autonomous-agent Long-running autonomous agent built at Lythe Labs, with persistent memory, sandboxed code execution, and a resumable control-plane API Python, FastAPI, SQLite, React
agent-testing-suite Local-first observability and evaluation framework for AI agents built at Lythe Labs: execution tracing, live dashboard, multi-model support TypeScript, Next.js
ai-cache Automatic caching layer for LLM API responses across OpenAI, Gemini, and Anthropic Python, LLM APIs

📊 GitHub Stats:

Top Languages      GitHub Activity

Pinned Loading

  1. crucible crucible Public

    Multi-agent collaboration engine with persistent, searchable, self-compacting memory. Agents research, critique and adjudicate over a shared SQLite memory with hybrid BM25 + embedding recall. Zero-…

    Python 2

  2. lythelab/voice-agent-starter-kit lythelab/voice-agent-starter-kit Public

    A production-ready voice agent template that feels like a real phone call

    Python 8 1

  3. athar athar Public

    Local-first, time-travel debugger for multi-agent LLM systems. Record execution DAGs into SQLite, scrub visual timelines, and fork-and-replay from any step without re-running earlier nodes. Zero cl…

    Python 1

  4. ai-cache ai-cache Public

    Transparent, cost-aware caching for LLM API responses. Cut spend and latency on repeated OpenAI, Anthropic, and Gemini calls without changing your code.

    Python 1

  5. stock-market-predictor stock-market-predictor Public

    Stock price prediction app using Streamlit and PyTorch LSTM, featuring real-time data from Yahoo Finance, customizable inputs, and interactive visualizations.

    Python

  6. mailgraph mailgraph Public

    Local-first email intelligence and safe cleanup desktop application. Interactive relationship graph, sender management, full-text search, and time-travel mailbox analytics.

    Python