Skip to content
View omid-sar's full-sized avatar
🎯
Focusing
🎯
Focusing

Block or report omid-sar

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.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. 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
omid-sar/README.md

I am the Founder of Memory Bridge and a Senior ML Engineer based in Ottawa, Canada πŸ‡¨πŸ‡¦.

My work sits at the intersection of System Architecture, Product Strategy, and Applied AI. I build privacy-first memory layers that give LLMs long-term context, bridging the gap between raw model capability and reliable, economic business utility.

Alongside product work, I’m active in AI safety learning and community circles (including BlueDot Impact), where I focus on practical questions around trustworthy systems, alignment constraints, privacy, and real-world deployment trade-offs.


πŸ”­ Strategic & Technical Focus

🧠 Context-Aware AI Agents
Architecting provider-agnostic memory vaults and retrieval systems. My focus is on solving the "amnesia problem" in LLMs to enable persistent, personalized user experiences while maintaining strict privacy boundaries.
πŸ›  Production Engineering
End-to-end engineering for SOTA open-weights (Llama 3.x, Mistral, Qwen). My work goes beyond standard SFT, leveraging Preference Alignment (DPO, PPO) and advanced PEFT strategies to steer model behaviors. I optimize for high-throughput serving (vLLM, TGI) and economic scaling on AWS/Kubernetes.
πŸ›‘οΈ AI Safety & Governance
Researching alignment constraints, interpretability, and the societal impact of autonomous agents. I work to translate theoretical safety principles into concrete engineering guardrails.
πŸš€ Product & Strategy
Aligning technical architecture with business value. I leverage techno-economic analysis to build AI products that are not just technically novel, but commercially viable and user-centric.

Building the memory layer for the next generation of AI.

Pinned Loading

  1. End_to_End_Text_Summarizer_AWS End_to_End_Text_Summarizer_AWS Public

    Jupyter Notebook

  2. End_to_End_Transformer_En_Fa End_to_End_Transformer_En_Fa Public

    Jupyter Notebook 2

  3. Finetuning_Llama2_7b_in_French Finetuning_Llama2_7b_in_French Public

    Jupyter Notebook 2

  4. YouTube-Assistant-LangChain YouTube-Assistant-LangChain Public

    Python

  5. LendingClub.com-Omid LendingClub.com-Omid Public

    Jupyter Notebook

  6. MNIST-Data--Digit-Recognizer- MNIST-Data--Digit-Recognizer- Public

    Jupyter Notebook