Building AI
that works
in the wild.
Hi, I'm Inder. I build intelligent systems, study what makes them reliable, and share the engineering lessons along the way — from recommendation systems to agentic AI.
Systems > demos
Where research meets
engineering.
Recommendation systems
Retrieval, ranking, embeddings, and building fast personalization pipelines.
Explore this area ↗02 / Agentic AI↗Reliable AI agents
Orchestration, tool use, retries, evaluations, and systems that recover gracefully.
Explore this area ↗03 / ML Infrastructure↗Model training & serving
Scaling training, inference, and monitoring for practical LLM applications.
Explore this area ↗Always building.
Always learning.
Applied research, machine learning infrastructure, and real products. A short version of the journey so far.
View professional profile ↗Senior Machine Learning Engineer
MetaRecommendation and advertising ranking systems at production scale.
Senior Applied Scientist
OmnekyAgentic systems, reinforcement learning, and generative AI.
Applied Scientist / Data Scientist
OmnekyMultimodal systems, NLP, optimization, and applied machine learning.
Try the agent explorer.
From signal to
safe action.
Click a stage to explore a simple mental model for reliable AI agents. The diagram and status update as you move through the workflow.
Read the system state before making a decision.
Notes from the field.
Short technical explainers below are clearly labeled samples, ready to be replaced with my original published writing.
Reliable AI agents: what happens after a tool timeout?
A practical mental model for making agents resilient when the outcome of a tool call is uncertain.
Recommendation systems, explained from retrieval to ranking
Why large-scale recommenders use several stages instead of scoring every item with one giant model.
Curious minds build
better systems.
Interested in machine learning, research, or the design decisions behind reliable AI?