Artasam Bin Rashid

About

I build AI systems that actually make it to production — not just notebooks that work once. Most "AI engineers" can prompt a model. I design the systems around it: multi-agent architectures, retrieval pipelines, and the MLOps plumbing that keeps them running after the demo ends. What I've shipped: 🔹 A 7-node LangGraph StateGraph powering a multi-agent system — orchestrating specialized agents instead of one overloaded prompt 🔹 An end-to-end MLOps pipeline (ingestion → training → CI/CD retraining) using GitHub Actions + Hopsworks Feature Store — the kind of infra that lets models improve themselves instead of going stale 🔹 A fine-tuned Stable Diffusion Inpainting model for domain-specific generative vision 🔹 Production-grade RAG pipelines — hybrid retrieval, reranking, hallucination-guarded prompting — built to survive real queries, not benchmark ones Stack I work in daily: LangChain / LangGraph · FastAPI · MLflow · Hopsworks · Chroma · Sentence-BERT embeddings · GitHub Actions · Python What ties it together: I don't see "the model" and "the system around the model" as separate problems. An agent is only as good as its retrieval, and retrieval is only as good as the pipeline retraining it — so I build across the whole stack, not just the flashy part. Currently deepening my work in agentic RAG and autonomous orchestration, and exploring graduate study to push further into LLM systems research. Always up for talking shop on RAG architecture, agent design, or MLOps — feel free to connect.

What I'm looking for

Remote AI/ML Engineering roles in LLM-powered agentic systems, RAG pipelines, and MLOps. I build production multi-agent architectures (LangGraph, LangChain) and own full ML lifecycles — data pipelines through CI/CD retraining (Git/GitHub, Hopsworks). Open to Applied AI, NLP, or Deep Learning Engineer roles focused on real deployment, not prototypes. Especially interested in teams working on agentic RAG, autonomous orchestration, or scalable LLM infrastructure.

Experience

Data Science Intern

10Pearls Pakistan

Dec 2025 – Feb 2026

• Architected production AQI forecasting system on Open-Meteo real-time API; XGBoost achieved RMSE 0.66, R2 0.99, out- performing Random Forest by 20%. • Integrated Hopsworks Feature Store for feature versioning and model registry; GitHub Actions CI/CD eliminated manual retraining overhead across 60+ consecutive operational days. • Engineered time-based, rolling-statistics, and lag features across a fully serverless stack with hourly ingestion and daily retraining pipelines.

XGBoostHopsworks Feature StoreGitHub

AI/ML Intern

DAM TECHHUB

Jul 2025 – Sep 2025

• Built AI-powered Resume Analyser: Logistic Regression + TF-IDF classifier achieving 87% accuracy across 24 job categories on a 2,400-resume dataset; Streamlit deployment with sub-1s inference latency. • Integrated Hugging Face S-BERT for cosine semantic similarity matching — 23% precision improvement over keyword-only baseline. • Extended platform with ATS simulation across 6 systems (Taleo, Workday, Greenhouse, iCIMS, Lever, Generic) and Groq Llama 3.3 70B AI bullet-polish rewriter.

TF-IDFHugging Face S-BERTStreamlit

Education

Certifications

Skills