Discover
Map the workflow, users, constraints, data sources, and risks before writing code.
I build AI-integrated products, research software, and infrastructure tools across the full stack: from product UX and APIs to model workflows, deployment, and hardware-adjacent prototypes.
Document ingestion, grounded responses, human review, and deployment-ready product surfaces.
Learning, focus, mobile UX, browser workflows, and polished user-facing systems.
Research computing support, retrieval, documentation, and operational clarity.
I am Amir Thapa Magar, a software engineer with a computer science research background and a product-builder mindset. My work spans AI workflows, full-stack systems, HPC and research infrastructure, mobile foundations, and applied machine learning.
The current focus is clear: build useful AI-integrated software with credible architecture, thoughtful UX, privacy-aware workflows, and deployment paths that can survive real users.
The strongest projects sit first: AI workflows, HPC tooling, document intelligence, language learning, focus software, and research systems from my personal engineering practice.
Caminode is the software studio side of my engineering practice: focused on SaaS, AI workflows, productivity software, technical tools, and products that respect real deployment constraints.
AI IntegrationFull-Stack EngineeringHPC & InfrastructureProduct ThinkingThe work is not positioned as a generic skill list. Each capability is framed around product evidence, architecture, privacy, deployment, and human review.
Map the workflow, users, constraints, data sources, and risks before writing code.
Turn the system into a usable product surface with clear states, review points, and guardrails.
Ship typed interfaces, APIs, AI workflows, data handling, and deployment-ready code.
Validate with real inputs, source evidence, monitoring needs, and paths for iteration.
Practical AI features with agents, document processing, source-grounded responses, review flows, and privacy controls.
From product interface to backend workflow, data model, deployment, and iteration-ready architecture.
Research-computing workflows, documentation systems, Linux-oriented operations, and cluster support tooling.
Model training, evaluation, inference APIs, sensor data, image classification, and on-device ML prototypes.
Native and cross-platform application delivery for polished mobile experiences, prototypes, and production workflows.
Sensor workflows, embedded foundations, data logging, microcontroller prototypes, and hardware/software integration.
The mobile and embedded work stays visible as career foundation, but it no longer dominates the primary story.
HPC support, applied machine learning, document intelligence, embedded foundations, and computer-vision experiments that support the current AI and infrastructure positioning.
A compact view of the training themes that support the portfolio: applied AI, full-stack systems, infrastructure, mobile, and embedded engineering.
Model training, evaluation, computer vision, sensor data, and practical inference workflows.
Typed frontends, backend services, APIs, deployment, privacy-aware document workflows, and SaaS product UX.
Docker, Linux operations, CI workflows, HPC documentation, and deployment-ready engineering habits.
iOS/Android delivery, hardware-adjacent data logging, sensors, and on-device ML experimentation.
This is the practical engagement layer of EngineeringByAmir: software, AI, infrastructure, and hardware-adjacent builds with clear technical ownership.
Document intelligence, LLM workflows, review systems, source-grounded answers, privacy scrubbing, and useful internal tools.
SaaS MVPs, dashboards, portals, admin tools, backend APIs, and polished interfaces that can grow from prototype to product.
Native and cross-platform mobile apps, product prototypes, app-store-oriented features, and mobile-first UX flows.
Research-computing portals, cluster support tooling, scientific workflows, documentation systems, and Linux-oriented operations.
Sensor data logging, microcontroller-oriented prototypes, hardware/software integration, and data pipelines for applied ML.
Currently available for hiring and freelance work.