Skip to content
View Shaon2221's full-sized avatar
🏢
Working
🏢
Working

Highlights

  • Pro

Block or report Shaon2221

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.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

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

Report abuse
Shaon2221/README.md

⚙️ Shaon Sikder — AI Engineer (GenAI & MLOps)

I help teams ship production‑grade LLM features fast — and safely. From fine‑tuning and agents to advanced RAG and AWS/K8s deployments, I build systems that move metrics (engagement up, errors down, time‑to‑value short).

• Portfolio: https://shaon2221.github.io • LinkedIn: https://www.linkedin.com/in/shaon2221 • Email: shaonsikder.ewu@gmail.com • YouTube: https://www.youtube.com/@DataDiscoveryBD

I care about fast feedback loops, clear communication, and measurable impact.


How I can help

  • Design & ship LLM copilots/agents (LangGraph, LangChain, Langsmith, ADK, Pydantic AI, CrewAI, n8n) that actually close loops
  • Fine‑tune and optimize models (GRPO/DPO, LoRA/PEFT, quantization) for cost and latency
  • Build advanced RAG with evals and guardrails (ChromaDB/Milvus/Pinecone + tracing)
  • Production MLOps: Docker/Kubernetes, CI/CD (GitHub Actions/Jenkins), MLflow/DVC on AWS

Proof in outcomes

  • +20% user engagement with a fine‑tuned Llama chatbot (Easital)
  • −35% OCR errors via multimodal llm improvements (Easital)
  • +10% lift vs betting odds with esports prediction (Opus)
  • −20% simulated risk using social network analysis (Piana)

Recent work

  • Short Circuit Science (2024–Now): medical imaging & sports analytics; multi‑agent copilots; advanced RAG; API backends; AWS/K8s
  • Actionboard.ai / Deep Data Lab (Contract): MCP servers for real‑time intent, knowledge graphs; robust MLOps
  • Delineate (Contract): AI agents for drug discovery; on‑prem LLM search for PK/PD data

Tooling I trust

Python, PyTorch, TensorFlow, scikit‑learn, FastAPI/Flask/Django, SQL/PostgreSQL, MongoDB, Redis, Docker, Kubernetes, AWS (EC2, S3, SageMaker, Lambda), MLflow, DVC

Also teaching & building in public

  • Instructor: Generative AI (Horizon Solutions, 2024) • Fundamentals of AI (Skill Hunt, 2021)
  • Creator: Data Discovery — practical GenAI tutorials and experiments

Research (ongoing)

Transfer learning for bone marrow classification • Thermal comfort ML + XAI • Federated learning in smart buildings

Education

BSc, Information & Communication Engineering — East West University (GPA 3.46/4.0)

— If you’re exploring LLMs/agents or need a production‑ready RAG/MLOps setup, let’s talk. 🚀

Pinned Loading

  1. Learning-and-Experimenting_Data-Science Learning-and-Experimenting_Data-Science Public

    Journey through data analysis

    Jupyter Notebook 28 14

  2. Walmart-SalesBot-and-SearchGPT Walmart-SalesBot-and-SearchGPT Public

    This project has been completed using large language model, langchain, prompt engineering, vector database etc.

    Jupyter Notebook 1

  3. Real-Time-Tweet-Analysis-on-US-Election-2020 Real-Time-Tweet-Analysis-on-US-Election-2020 Public

    An End-to-end project of NLP

    Jupyter Notebook 3

  4. Data-Science-Job-Analysis Data-Science-Job-Analysis Public

    Salary Prediction, Insight/trend extraction, Web-scraping from glassdoor.com (End-to-End Project)

    Jupyter Notebook 4 1

  5. Insights-Extraction-from-my-own-Facebook-Profile-Data Insights-Extraction-from-my-own-Facebook-Profile-Data Public

    Explored my own Facebook profile data!

    Jupyter Notebook 4

  6. Fake-News-Detection-using-LSTM Fake-News-Detection-using-LSTM Public

    Deep Learning model which can predict 99% accurate for given Dataset

    Jupyter Notebook 2 1