I'm Ajay Kumar Reddy KrishnareddyGari, a Creative Technologist and AI Engineer based in India with over three years of professional experience building intelligent systems, machine learning pipelines, and high-performance web applications. My work sits at the intersection of deep learning research and product engineering — from training generative adversarial networks to shipping production-grade React interfaces.
I specialize in machine learning, computer vision, generative adversarial networks (GANs), and full-stack web development. My engineering philosophy is grounded in building systems that are not only technically rigorous but also deliver exceptional user experiences. I believe the best AI products combine mathematical precision with thoughtful product design.
My work spans computer vision, natural language processing, healthcare AI, autonomous supply chain systems, and civic technology. Each project is built to solve a real problem with a bias toward clean architecture and exceptional user experience. I've led teams building AI-first products that have been deployed in healthcare, logistics, and civic engagement contexts.
Beyond building AI pipelines, I invest heavily in developer tooling and workflow automation. I'm a strong advocate for open-source development; most of my work lives publicly on GitHub. I'm also professionally connected on LinkedIn.
When not writing code, I explore generative adversarial network architectures, contribute to open-source computer vision tooling, and publish experimental interfaces. If you're interested in collaborating or have an engineering challenge worth solving, reach out via the contact section below.
A curated archive of design-led engineering and experimental digital solutions. Each project represents a significant engineering challenge solved with cutting-edge AI and full-stack development techniques. Browse the full selected works section or the extended project archive for the complete list.
A GAN-based face transformation suite enabling controllable emotion editing, identity morphing, and high-fidelity sketch generation. Built using PyTorch and drawing from foundational research in Generative Adversarial Networks (Goodfellow et al., 2014), this system provides state-of-the-art facial transformation capabilities. The architecture uses a conditional GAN framework with custom loss functions for perceptual similarity and identity preservation.
An AI-powered patient triage system that analyzes symptoms, intelligently prioritizes cases, and dynamically optimizes doctor queues for efficient clinical decision-making. Medyphas AI uses natural language processing to parse patient symptom descriptions and a multi-factor scoring algorithm to determine case urgency. The system integrates with hospital management workflows to reduce wait times and improve patient outcomes.
An AI-powered frontend generation pipeline that uses retrieval-augmented generation to produce structured, consistent, and pixel-aligned UI layouts. The system combines a vector database of design patterns with a large language model to generate production-ready React components from natural language descriptions. This pipeline leverages techniques from the RAG research by Lewis et al. (2020) and has been used to generate complete application UIs from single prompts.
A context-aware weather interface that transforms forecast data into actionable insights for smarter, real-time decision-making. Tempo uses adaptive data visualization techniques to present atmospheric data in a human-readable format, integrating with multiple weather data APIs to provide hyperlocal forecasts and trend analysis.
An autonomous logistics intelligence and predictive monitoring system built to handle real-time supply chain disruptions. Sentinal-X uses time-series machine learning models to predict inventory shortfalls and automatically route around supply chain failures.
A civic intelligence platform leveraging Google Vertex AI (Gemini 1.5 Pro) for manifesto summarization and trial electronic voting machine simulation. V2I makes political information accessible and verifiable, supporting informed democratic participation. Live at vote2india.vercel.app.
A curated set of commands, workflows, and systems used daily in building AI and web projects. This toolkit includes Git version control workflows, Bun and npm package management, Python environment management with uv, and architectural cheatsheets for common patterns including REST APIs, GraphQL, OAuth2, and JWT authentication flows. Browse the interactive Developer Toolkit section for command references and cheatsheets.
I work across the full stack with a primary focus on Python, PyTorch, and TensorFlow for machine learning, and React, TypeScript, and Node.js for web development. For backend APIs I use FastAPI and PostgreSQL. On mobile, I've shipped Android apps with Kotlin. I also use Figma for design and have deep experience with the PyTorch ecosystem and the broader Hugging Face open-source ML community.
My AI/ML expertise includes: supervised and unsupervised learning, convolutional neural networks (CNNs), recurrent neural networks (RNNs), transformer architectures, generative adversarial networks (GANs), and retrieval-augmented generation (RAG). On the engineering side I have production experience with React, Next.js, TypeScript, Python, FastAPI, Node.js, PostgreSQL, MongoDB, Docker, and cloud platforms including Vercel and Google Cloud.
Ajay Kumar Reddy KrishnareddyGari specializes in AI engineering, machine learning, generative adversarial networks (GANs), computer vision, and full-stack web development using React, TypeScript, Python, and FastAPI. He has over three years of professional experience building production AI systems and full-stack applications.
Ajay has built the AI Face Transformation Suite (GAN-based emotion and identity morphing), Medyphas AI (intelligent patient triage), an AI UI Generation Pipeline using RAG, Tempo (weather intelligence interface), Narrative Shield (misinformation detection), Sentinal-X (autonomous supply chain intelligence), and V2I Vote to India (civic technology with Google Gemini AI). The full project archive is available in the portfolio section.
Ajay works with Python, PyTorch, and TensorFlow for machine learning; React, TypeScript, and Node.js for web development; FastAPI and PostgreSQL for backend APIs; and Kotlin for Android development. He uses Figma for design and has deep experience with the PyTorch ecosystem.
The AI Face Transformation Suite is a GAN-based system enabling controllable emotion editing, identity morphing, and high-fidelity sketch generation from facial images. It was built using PyTorch and draws from research in Generative Adversarial Networks (Goodfellow et al., 2014). See the selected works section for more details.
Yes. Ajay is open to both freelance project engagements and full-time engineering roles, particularly those involving AI systems, computer vision, or full-stack product development. Connect directly on LinkedIn or reach out via the contact section.
All public projects and experiments are available on github.com/ajaykumarreddy-k. The repository covers GAN training scripts, full-stack web apps, Android applications, and civic technology tools. Browse the archive section for annotated project summaries.
Reach Ajay via email at ajaykumarreddykrishnareddygari@gmail.com, connect on LinkedIn, or explore his open-source work on GitHub.