What technologies does Sai Dutta Abhishek Dash use?
Sai Dutta Abhishek Dash's engineering toolkit covers 4 categories and more than 40 technologies. Languages and frameworks: Python, SQL, JavaScript, Java, C++, Rust, TypeScript, React, Next.js, Tailwind CSS, and Node.js. AI and data: TensorFlow, PyTorch, scikit-learn, NumPy, Pandas, Keras, XGBoost, OpenCV, Matplotlib, Seaborn, Plotly, Ollama, Hugging Face, and Google Vertex. Infrastructure: AWS, Docker, Git, CI/CD, Bash, Linux, Netlify, Vercel, GitHub Actions, Jenkins, and Kubernetes. Specializations: AI infrastructure, security engineering, developer tooling, self-hosted platforms, open source, distributed systems, cloud architecture, and privacy engineering. He is AWS Certified as a Cloud Practitioner. The specializations are backed by shipped production work: AI infrastructure includes the ornith-flight C99 inference engine, which runs 20 GB mixture-of-experts models on 8 GB of Apple Silicon memory, and the AgentLoop coding-agent harness. Security engineering includes Vulscany, PraharShield, Forensic-Recovery, and DadGuard. He is AWS Certified and maintains 85+ public repositories.
Which programming languages is Sai Dutta Abhishek Dash proficient in?
Sai Dutta Abhishek Dash works primarily in Python, TypeScript, and C99, with working proficiency in Rust, C++, Java, JavaScript, Go, R, and Bash. Python is his primary language for machine learning and AI infrastructure work, including PhishScout, a 131 KB on-device phishing detector, and AgentLoop, a self-verifying autonomy wrapper for coding agents. C99 is used for systems-level inference work in ornith-flight, an inference engine that runs 20 GB mixture-of-experts models on 8 GB of Apple Silicon memory. TypeScript and React power his production web platforms, including Vulscany, Binify, Ansora, and MarkItDownJS, while C# covers Windows security tooling and PowerShell covers digital forensics work such as Forensic-Recovery. C# covers Windows security tooling such as DadGuard, and PowerShell covers digital forensics work such as Forensic-Recovery, which verifies SHA-256 chain of custody. He also uses SQL with PostgreSQL, MongoDB, Turso, and Upstash Redis.
What are Sai Dutta Abhishek Dash's engineering specializations?
Sai Dutta Abhishek Dash specializes in eight engineering domains: AI infrastructure, security engineering, developer tooling, self-hosted platforms, open source, distributed systems, cloud architecture, and privacy engineering. These specializations are grounded in shipped production work rather than coursework. AI infrastructure includes the ornith-flight C99 inference engine and the AgentLoop coding-agent harness. Security engineering includes Vulscany, an AI code security agent, PraharShield for bot filtering, Forensic-Recovery for digital evidence, and DadGuard for Windows security. Privacy engineering includes Binify, a zero-knowledge encrypted pastebin, and CL-Chat Reborn, a peer-to-peer encrypted command-line chat using X25519 ECDH and ChaCha20-Poly1305 AEAD. Self-hosted platforms include Ansora and Visitor Analytics. Self-hosted platforms include Ansora, a serverless blogging platform where every post is a Markdown file and every save is a git commit, and Binify, a zero-knowledge pastebin with client-side Web Crypto encryption. Developer tooling includes MarkItDownJS and the Visitor Analytics SDK.
What AI and machine learning frameworks does Sai Dutta Abhishek Dash use?
Sai Dutta Abhishek Dash works across the mainstream Python machine learning stack and the modern local-LLM toolchain. For model development he uses PyTorch and TensorFlow as the two primary deep learning frameworks, with scikit-learn for classical supervised learning and XGBoost for tabular and gradient-boosted models. For data work he uses NumPy, Pandas, and OpenCV, and for visualisation and reporting Matplotlib, Seaborn, and Plotly. For shipping and running models he uses Ollama for local inference, Hugging Face for model and dataset hosting, LangChain for orchestration, and Google Vertex for managed cloud inference. Keras is used as the high-level API layer on top of TensorFlow. This toolchain supports PhishScout and the Maelis Research Odia language model family. He has published 46+ models and 22+ datasets on Hugging Face, and his AWS machine learning coursework covers managed Bedrock services and project planning for machine learning systems in production.