Blogs

How does distributed tracing work? Follow trace IDs, spans and W3C headers across services, with interactive diagrams, sampling trade-offs and async context.

How should you handle AI cheating in technical interviews? Set clear tool rules, assess real understanding and separate skill evidence from identity fraud.

How do you add Langfuse to Hermes Agent? Enable the bundled plugin, configure project keys and storage, and inspect traces, cost estimates and capture settings.

How do prompt, context, harness, loop and evaluation engineering differ? Compare five AI agent disciplines, failure signals and practical checks for production.

Who am I if my body, experiences and habits keep changing? I explore identity through rivers, Buddhist not-self and Ramana Maharshi, without a final answer.

How do you verify AI subagent research? My Odia-resource failure, corrected dataset claim and practical checks for model IDs, paper titles and tool evidence.

Why did PyConf Hyderabad 2026 feel like a homecoming? My speaker experience, audience feedback and volunteering history brought me back to the Python community.

How can you keep up with AI without chasing every release? My February 2026 weekend story, what METR’s task horizons mean and how I choose a learning lane.

How is AI used in the adult industry? I explore open models, image licences, hardware protocols, consent and the limits of local privacy and low latency.

How can you make consistent blog covers? My CoverView fork adds layouts, color palettes and export fixes, with lessons on React state and browser rendering.

Why did my RSVP speed reader fail to fit my workflow? A prototype story about reading comprehension, skimming, AI prompts and knowing when to stop building.

How do you profile specific Python functions with tprof? Compare json and orjson, understand elapsed timing and cProfile scope and keep benchmarks reproducible.

How do you pass the OpenTelemetry Certified Associate exam? My 73% to 88% journey, current OTCA format, fees, curriculum and a practical preparation plan.

What stood out in my 2025? Personal year review of teaching, judging a GenAI hackathon, learning to drive, self-hosting, family history and lessons from books.

Is self-hosting cheaper than managed apps? My move from Heroku to Hetzner and a Proxmox home lab explains Coolify, historical costs and recovery trade-offs.

How can you influence colleagues without authority? Compare role, relational and expertise power, build useful relationships and share knowledge responsibly.

How has human-digital communication evolved? Compare handwriting, typing, dictation and brain-computer interfaces without confusing milestones with predictions.

How do you govern AI responsibly? An executive playbook covering EU AI Act dates, India’s DPDP rollout, NIST risk management and a practical 90-day action plan.

How do you monitor LLM applications? Use OpenTelemetry traces, Prometheus queries and evaluation evidence to track latency, quality and estimated AI cost.

What are the hidden costs of LLM-as-a-Judge? Calibrate biased judges, choose task-specific rubrics and control evaluation spend with sampling and human review.

What is RSS and how do you use it? Follow RSS and Atom feeds in a reader, explore my original subscriptions and check which engineering blog links still work.

How do RAG approaches differ? Compare Naive RAG, query refinement, HyDE, GraphRAG and LightRAG, with practical trade-offs, evaluation and a Python example.

How did I judge a GenAI hackathon? My HydPyHack 2025 experience covers sponsor use cases, success criteria, a weighted scoring sheet and choosing the winners.

How do you create Python packages using Rust? Build a PyO3 extension with maturin, test its API and understand wheels, ABI compatibility and GIL behavior.