Hi, I'm
Researching and evaluating frontier AI: forming hypotheses, validating with evidence, and translating what works into production.
I investigate emerging AI approaches, run experiments to separate signal from hype, and turn what works into real-world systems, not just prototypes.
Currently building AI Concepts Playground, a hands-on way to understand AI concepts through experimentation.
11+
Years Experience
6+
Technical Papers
80+
Blogs Published
Applied AI & Emerging Digital Technology Researcher who investigates frontier AI, validates it with evidence, and translates research into production.
Over 11+ years I've moved from full-stack engineering into researching and evaluating emerging technologies, and today I treat frontier AI as a research problem before an engineering one.
I'm an Integrated Emerging Digital Tech Analyst at Shell. I research emerging technologies, evaluate them through proofs-of-concept, and translate findings into reports, recommendations, and production-ready systems, spanning Generative AI, Agentic AI, HPC, Digital Twins, and Spatial Computing.
Before committing to an approach, I evaluate new AI methods, benchmark alternatives, and choose architectures based on empirical evidence, measuring value, performance, and cost rather than following hype.
I publish what I learn, open-source the tools I build, and share findings with the wider community. I believe good research isn't about using more technology; it's about rigorously understanding what actually works and why.
I approach emerging AI as a research problem first: forming hypotheses, running experiments, benchmarking approaches, and publishing what I learn. The result is 6+ technical papers and dozens of proofs-of-concept that shaped enterprise technology roadmaps.
6+
Technical Papers
20+
Proofs-of-Concept
80+
Articles Published
Research paper presented at IIM Bangalore
Developed a 3D synthetic environment in which robotic arms learned object grasping and pick-and-place tasks using vision and sensor feedback. The trained policies were designed for deployment to ROS-based robotic systems, bridging simulation-based reinforcement learning with real-world robotics.
Investigated retrieval-augmented generation, semantic search, and multi-agent workflows for technical knowledge management, evaluating chunking, embedding, and grounding strategies to improve answer accuracy and traceability.
Experimented with parameter-efficient fine-tuning (LoRA, QLoRA) and systematic evaluation to adapt foundation models to domain data, tracking experiments and comparing cost, quality, and latency trade-offs with MLflow.
Studied how High-Performance Computing can parallelize extraction, preprocessing, embedding generation, and vector indexing, measuring throughput gains for enterprise-scale document workloads.
Evaluated Digital Twin, OpenUSD, and spatial-computing approaches through technical research and internal publications, assessing feasibility and impact to guide enterprise innovation roadmaps.
A hands-on way to build real AI intuition, one interactive lesson at a time.
Interactive experiments that simplify LLMs, RAG, vector search, AI agents, and other modern AI concepts through visual, hands-on learning.
Explore topics like
Why Learn Here
Free • No Sign-up • Open to Everyone
These projects weren't built as demos. They were created to solve real engineering challenges encountered while building AI products. Each one is open source, production-ready, and designed to be reused, extended, and improved by the community.
A production-ready, provider-agnostic AI chat framework that lets teams switch LLM providers without rebuilding the application.
Instead of coupling the user interface to a single model provider, ChatUI separates the frontend from the AI backend through a configurable provider layer. Whether you're using Azure OpenAI, OpenAI, Ollama, or another compatible API, the same application can switch providers without changing the UI.
Why it matters
Turn "reprocess everything" into "process the delta."
File Tracker is an incremental processing framework that detects exactly what has changed between runs, eliminating unnecessary work in AI and data pipelines. Instead of reprocessing every document, it identifies new, modified, and unchanged files so only the delta moves through the pipeline.
Why it matters
Event-driven ingestion, so the system reacts to reality.
FolderWatcher automatically detects file creations, modifications, deletions, and moves, triggering downstream workflows the moment changes occur. Instead of relying on scheduled scans, it enables event-driven ingestion so pipelines stay up to date while minimizing unnecessary processing.
Why it matters
Right-size vector search: start on a file, scale to a cluster.
milvuslite-kit simplifies semantic search by letting you define collections, schemas, and indexes declaratively in YAML while running on Milvus Lite. Start with a lightweight, infrastructure-free setup for development, then transition to Milvus Standalone or Distributed without changing your data model.
Why it matters
Ask questions about a folder of Markdown files - and get answers grounded in your documents.
No database, no embeddings, no vendor lock-in. Instead of chunking files into a vector index, it organizes knowledge with the Open Knowledge Format (OKF) and lets an LLM browse the folder like a new employee - reading indexes, metadata, and sections before answering. Works with OpenAI, Azure AI Foundry, or fully offline.
Why it matters
Articles on software engineering, cloud, AI, XR, and emerging technology, published on Medium.
A decade of building AI, cloud, and immersive systems, from research and prototypes to production.
Shell India Markets Pvt Ltd
via L&T Technology Services (2018–2023) · via Randstad India Pvt Ltd (2023–2025) · Direct (2025–Present)
Generated an estimated USD 18 million uplift through improved safety and operational efficiency of oil and gas assets by leveraging the right emerging digital technologies, including AI, Generative AI, HPC, and spatial computing.
Key Highlights
ASM Technologies Ltd
Axinovate Technologies LLP
A cross-domain toolkit spanning the full stack, from frontend interfaces to cloud infrastructure and emerging technologies.
User Interfaces & Experiences
APIs & Server-Side Logic
Infrastructure & Deployment
LLMs, RAG & AI Services
Next-Gen & Immersive Domains
Ela is the epitome of intellectual curiosity and humility backed by deep expertise in the XR tech stack. Very rarely does one come across someone with a sense of humor, a willingness to help, roll their sleeves and get stuck in and deliver in the face of crazy deadlines, Ela would be the first one to put his hand up when you have a task that requires all of the above, would do it with a smile, learn from work and end up excelling in whatever he takes on. True asset to the team and does everything he can to share his learnings to a larger audience with the intent to create a community around technology and business impact.
True pleasure to work with and would have him in my team in a heartbeat. Highly recommended!!!
Elanthirayan Madhavan was part of Shell's XR & Immersive Analytics team that I managed in 2022, and I couldn't have asked for a more capable colleague. Balancing studies alongside his contributions, Elan was one of the team's most dependable contributors. His deep technical knowledge regarding VR and AR was always matched by his willingness to assist others. He never hesitated to share his expertise or lend a hand.
Elan was instrumental in advancing several VR projects. Whether he was refining a technical VR simulation or troubleshooting last-minute bugs, he tackled each task thoughtfully and with precision. He has that rare blend of technical smarts, creative problem-solving, and calm under pressure, and he does it all with a warm sense of humor.
But beyond the work itself, what truly stands out is how he fosters collaboration. Elan brings positive energy to meetings and offers feedback in a way that's both helpful and uplifting.
If you're seeking someone who can drive immersive experiences forward while being a steady, likable teammate, Elan is that person. I'd gladly work with him again and have no doubt any team would be fortunate to bring him on board.
Ela was part of my team at Shell, and he quickly stood out as a fast learner with a strong passion for technology. Ela is naturally curious, proactive, hardworking and always eager to explore new tools and ideas. While working in XR, he was self-motivated to complete a Masters in Data Science and broaden his skills. He brings a wide range of technical knowledge, from XR to web technologies and cloud computing, and ability to interface with the business. Ela also worked effectively with third-party partners. Ela is a driven, well-rounded professional, and I'm confident he'll continue to make a meaningful impact wherever he goes.
I had the pleasure of working with Ela during our time at Shell, and one thing that always stood out was his genuine curiosity and fearless approach to trying new things. Ela is the kind of person who doesn't just keep up with emerging technology; he actively seeks it out, dives in, and brings others along for the ride.
We collaborated on several initiatives, including delivering technical sessions at Universiti of Technology Brunei through the Shell.ai program. Teaching alongside Ela was a great experience; he has a natural talent for simplifying complex topics and engaging audiences with clarity and energy.
He also represented our team at global events, where he spoke passionately about the evolution of emerging technologies and the future of tech within Shell. One standout moment was when he delivered a live demo on XR technology, and not only did he explain it clearly, but he made it truly exciting and approachable for a broad audience.
Ela brings contagious energy and positivity to every team he's part of. He's collaborative, eager to learn, and never afraid to step into uncharted territory. Working with him was genuinely inspiring, and I'm excited to see the impact he'll continue to make in the tech space.
A few things I enjoy when I'm not in front of a screen, or sometimes while I am.
Open to new opportunities, collaborations, and interesting conversations across software engineering, AI, cloud, and emerging technology.