Skip to content
View NechamaKrashinski's full-sized avatar

Achievements

Achievement: Pull Shark

Achievements

Achievement: Pull Shark

Block or report NechamaKrashinski

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 is supported. 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
NechamaKrashinski/README.md

Nechama Krashinski

Software Engineer | Infrastructure, Full Stack & AI Agents

Analytical thinker and highly independent developer, with proven experience building complex, performance-critical end-to-end systems and AI-assisted workflows. Currently specializing in production-grade AI Agent architectures.

💻 Technical Expertise

Languages & Frameworks Python C# JavaScript C++ C Java

AI & Agentic Engineering LangChain LangGraph LlamaIndex PyTorch OpenCV

Backend & Frontend Architecture .NET Core Node.js React Redux Angular

Databases & Infrastructure SQL Server Redis MongoDB Docker Linux

🚀 Featured Projects & Engineering Impact

🧠 AI & Multi-Agent Architecture

  • Advanced AI Agent Development (Dovrot AI) Architecting and deploying production-grade AI agents with a focus on complex orchestration, real-world tools integration, and observability:
    • LangGraph Orchestration: Designing state-machine workflows with branching, loops, Human-in-the-Loop (HITL) checkpointers, and Multi-Agent patterns (ReAct, Plan-and-Execute).
    • Advanced RAG Pipelines: Implementing semantic search utilizing LlamaIndex, Pinecone vector database, and Cohere embeddings, complete with intelligent routing logic and fallback strategies.
    • Model Context Protocol (MCP): Building MCP Servers to grant LLMs robust control over external environments, including browser automation via Playwright.
    • Evaluation & Tracing: Utilizing LangSmith for deep observability, debugging, execution tracing, and prompt versioning.
  • Hyperconverged KV-Cache Offloading (Pliops - KamaTech Bootcamp)
    Architected a hyperconverged KV-cache offloading system to maximize LLM inference throughput. Provisioned high-throughput RAID0 storage for KVRocks, enabling a stable GPU HBM-to-SSD offloading pipeline with vLLM and LMCache.

⚡ Software Engineering

  • Kung Fu Chess - Real-Time Modular Python Engine
    Designed and implemented a real-time chess engine in Python featuring an advanced modular OOP architecture. Built a custom Pub-Sub infrastructure to decouple core game logic from OpenCV rendering.
  • OpenCV Contributor
    Contributed to the official open-source OpenCV library (PR #28404), refining error handling within core components by replacing raw assertions with descriptive runtime errors.
  • Consultancy Workflow & Scheduling Platform
    Implemented a secure, multi-role scheduling system using Node.js and React/Redux, featuring state-based workflow logic and role-based access control (RBAC) backed by a normalized SQL schema.

Pinned Loading

  1. Role-Guard Role-Guard Public

    C#

  2. Hyperconverged-KV-Cache-Offloading-for-Cost-Efficient-LLM-Inference Hyperconverged-KV-Cache-Offloading-for-Cost-Efficient-LLM-Inference Public

    Benchmarking Hyperconverged KV-Cache Offloading to SSD (vLLM, LMCache, KVRocks) for cost-efficient Llama-3-8B inference, demonstrating significant optimization potential.

    2 1

  3. KFChess KFChess Public

    A special chess game designed for millions of users

    Python

  4. NechamaKrashinski NechamaKrashinski Public

    Config files for my GitHub profile.

  5. wit wit Public

    A Python project that simulates local Git

    Python

  6. Consultancy-Workflow-Scheduling-Platform Consultancy-Workflow-Scheduling-Platform Public

    TypeScript 1