AI Engineer & Specialist | Architecting Production-Grade AI Systems & Alignment Research
I am an AI Engineer and Software Developer specializing in designing and deploying production-grade AI systems. My expertise spans Retrieval-Augmented Generation (RAG) architectures, edge-optimized Computer Vision, low-resource NLP, and Generative Modeling.
While my foundation is in full-stack engineering, my core focus is now on bridging applied AI with AI Alignment research—building systems that are not only powerful but interpretable, robust, and aligned with human intent. I am a self-synthesizer who loves to build, break, and learn from the process, and I actively write technical blogs to demystify complex AI concepts for the developer community.
🧠 Applied AI & Machine Learning: Architecting enterprise RAG systems (Hybrid Search, pgvector), Generative AI pipelines, Agentic workflows, and edge-deployed Computer Vision models. 🔬 AI Alignment & Research: Investigating model interpretability, mesa-optimization, and robust evaluation frameworks to ensure AI systems behave safely and predictably. ⚙️ Backend & Systems Engineering: Building scalable microservices and data-intensive backends using FastAPI, Django, PostgreSQL, and modern cloud infrastructure. ✍️ Technical Communication: Writing in-depth blog posts to break down complex AI concepts and sharing perspectives on AI ethics and alignment.
-
e-Gazette RAG System (Hybrid Document Intelligence):
Architected a production-grade hybrid retrieval and QA system for a 14GB+ corpus of government documents. Integrated a custom OCR pipeline (Dolphin VLM) for degraded historical texts, combining BM25 and pgvector for high-precision semantic search and automated citation. -
Edge-Optimized Computer Vision (EV Monitoring):
Developed real-time computer vision pipelines for Electric Vehicle monitoring, focusing on low-latency inference, model quantization, and deployment in resource-constrained edge environments. -
Face2Comic (cGAN Research):
Engineered a Pix2Pix conditional GAN to translate real human portraits into comic-style illustrations. Focused on structural preservation via U-Net architectures and PatchGAN discriminators for high-fidelity image-to-image translation. -
Generative AI & Agentic Workflows:
Building autonomous agents and generative pipelines leveraging LLMs, Model Context Protocol (MCP), and advanced prompt engineering to solve complex, multi-step reasoning tasks.
- Speaker & Subject Matter Expert: Delivered a session on "Open Source Contribution" at NIELIT Cybrella 2.0 (2025).
- Project Maintainer: Led open-source initiatives for SWOC 2025 and IWOC 2025, guiding contributors in MERN and AI-based projects.
- Active Contributor: Hacktoberfest 2022 & 2024, with a focus on AI/ML repositories and developer tooling.
🧠 AI/ML & Data Science:
⚙️ Backend & Frameworks:
🗄️ Databases & Vector Stores:
🎨 Frontend & Styling:
☁️ DevOps & Tools:
If you find my open-source projects or technical blogs helpful, you can support me here:
