Skip to content
Portfolio

Field notes

Stay ahead of the systems you already own.

Daily reading and a short learning stack, mapped to Kenyt: .NET 8, Angular, Azure, Elasticsearch, and AI products — plus the leadership work of a 9-person team.

Start learning nowSee the timeline
Daily slots
3Daily slots
Learning tracks
6Learning tracks
Live sources
6Live sources

01

Daily operating system

Not a firehose. Three slots that fit around shipping and a 9-person team.

Morning

15 min

Scan the live timeline

Two posts max. Prefer .NET / Azure / Elastic over generic AI hype. If it cannot change a decision this quarter, skip it.

Deep

40 min · 3× week

One architecture piece

System design, reliability, or AI evals. Take notes as if you will brief the team on Monday: problem, trade-off, what we would do.

Build

45 min · 3× week

Touch the learning track

Ship a tiny artifact: a spike, a doc, a dashboard query. Reading without a build does not stick at this level.

02

Start learning now

Sequenced for Kenyt: AI products on the existing C# / Angular / Azure / Elastic stack — not a career reset.

AI product engineering

Kenyt ships conversational AI. The job is evals, latency, cost, and retrieval quality — not demo prompts.

  1. 01Azure OpenAI + your own eval set (golden conversations, not vibes)
  2. 02RAG with Elasticsearch kNN / semantic search on real Kenyt corpora
  3. 03Tool-calling agents with hard timeouts, tracing, and fallbacks

Azure OpenAI docs ↗OpenAI evals ↗Elasticsearch vector search ↗

.NET 9 / Aspire / observability

You already moved 4.8 → 8. The next win is cloud-native ops: OpenTelemetry, Aspire, and fewer snowflake clusters.

  1. 01.NET Aspire dashboard against one Kenyt service
  2. 02OpenTelemetry traces from ASP.NET through Elasticsearch calls
  3. 03Native AOT / trimming only where it pays; ignore the rest

.NET blog ↗Aspire ↗ASP.NET Core ↗

Angular 16 → current

You already did v8 → v16. Signals, zoneless change detection, and new control flow are the next migration, not another rewrite.

  1. 01Signals + computed in one dashboard widget
  2. 02New control flow (@if / @for) on a high-traffic view
  3. 03Read the upgrade guide; plan v17+ like you planned v16

Angular blog ↗Update guide ↗Signals ↗

Search & cache at Kenyt scale

Elasticsearch and Redis already sit on the critical path. Vector search, query cost, and cache invalidation are the next reliability work.

  1. 01Profile the slowest ES queries; add slow-log + circuit breakers
  2. 02Hybrid lexical + kNN for conversational retrieval
  3. 03Redis as request cache with explicit invalidation, not TTL-only

Elasticsearch blog ↗Redis blog ↗Azure Cache for Redis ↗

Distributed systems you will be asked about

SaaS across clusters: outbox, sagas, multi-tenant isolation, and blast-radius. This is interview-and-incident material.

  1. 01Transactional outbox for one existing .NET write path
  2. 02Idempotent APIs and poison-message handling
  3. 03Multi-cluster rollout playbook (you already built the pipeline)

Azure Architecture Center ↗InfoQ Architecture ↗Microsoft Architecture ↗

Principal EM craft

Nine people, multiple clusters, OKRs. The gap is strategy, hiring bar, and making architecture decisions stick without becoming a bottleneck.

  1. 01One-page technical strategy per quarter, tied to OKRs
  2. 02Staff-engineer promotion criteria written down
  3. 03Incident review that produces a design change, not a slide

The Pragmatic Engineer ↗StaffEng ↗Microsoft Engineering ↗

03

Live reading timeline

Headlines from .NET, Azure, Angular, Elastic, InfoQ, and Microsoft DevBlogs. Regenerates every four hours.

Last refresh · 2 Sept, 14:29 IST

  1. 2 SeptCustomizing Angular Aria Tabs Quickly with Google Antigravity CLIAngular
  2. 1 SeptDefending against AI-fueled social engineeringElastic
  3. 31 AugIntroducing Azure Multicloud Interconnect for AWSAzure
  4. 31 AugInside Microsoft’s marketing team: Scaling expertise with AIAzure
  5. 31 AugPodcast: Scott Jenson on Evolving Desktop OS, Local-First, & Agentic UXInfoQ
  6. 29 AugPresentation: Architecting the Data Layer for AI Agents: from Transactional Systems to MCP and Semantic ModelsInfoQ
  7. 29 AugCloudflare Workers Accept Inbound TCP, with gRPC the First Protocol on TopInfoQ
  8. 28 AugStyling Mastery and Declarative Form SubmissionsAngular
  9. 28 AugPresentation: From DVDs to Global Streaming: How Netflix’s Commerce Architecture Actually EvolvedInfoQ
  10. 28 AugArticle: Post-Quantum Cryptography in Spring Boot: Four Patterns You Can Ship This SprintInfoQ
  11. 27 AugManaged PostgreSQL vs. self-hosted PostgreSQL: Key benefits and trade-offsAzure
  12. 27 AugHow Uno Platform uses .NET, MCP, and AI to build high quality apps.NET
  13. 27 AugHow can agencies unify smart city data to improve public services?Elastic
  14. 26 AugThe Economics of Agent Optimization: Four ways to lower the costAzure
  15. 26 AugAWS Introduces Specification-Driven Composition for Flexible Data WorkflowsInfoQ
  16. 26 AugArticle: Beyond Offset Lag: Computing Time in Queue for Apache Hudi Data Lake Pipelines at Petabyte ScaleInfoQ
  17. 26 AugMachine vs. machine: The new reality of cybersecurity in ANZElastic
  18. 26 AugForging India's Digital Future: Elastic{ON} Comes to MumbaiElastic

Standing sources

.NET blog

Release cadence, ASP.NET, and runtime notes — your production stack.

Azure Architecture Center

Reference architectures for the SaaS / multi-cluster problems you already run.

Angular blog

Signals and the post-16 upgrade path before the next forced migration.

Elasticsearch engineering

Query cost, vector search, and relevance — the conversational-AI data plane.

Feeds: .NET · Azure · Angular · Elastic · InfoQ · Microsoft DevBlogs