Showing posts with label nrip nihalani. Show all posts
Showing posts with label nrip nihalani. Show all posts

Friday, July 24, 2026

From Coder to Orchestrator: Best Practices for Tech Leads Using Claude Code

The era of the Tech Lead as the "fastest coder in the room" is over.

"Beginners use Claude to avoid thinking. Intermediate users use Claude to think faster. Power users use Claude to think better."  

 

Today’s top engineering leaders aren't just writing code—they’re orchestrating AI agents that can generate it at superhuman speeds. The new core competency? Writing detailed prompts, reviewing massive amounts of code, and investing in shared tooling.

Claude Code has emerged as a powerhouse in this new landscape (crossing $1B in annualized revenue within six months of launch). But raw AI power without discipline is dangerous. Left unchecked, it can easily destroy a working codebase.

Here is a comprehensive guide to using Claude effectively across the entire software development lifecycle, synthesized from top engineering managers and the Claude Code team.

 

Part I: The Mindset Shift 

You Are the Tech Lead; Claude is Your Genius Junior The single most important shift is understanding your new role. Claude is your genius junior engineer: fast, brilliant, but requiring constant coaching and explicit rules. It can generate thousands of lines of code in minutes, but it needs clear specifications and architectural guardrails.

Think in Roles, Not Just Code Most developers open Claude and immediately ask it to build something. This is a mistake. Think of Claude in four roles: Architect, Tech Lead, Senior Engineer, and QA. Most people jump straight to Senior Engineer. A production-ready workflow requires Claude to act as the Architect and Tech Lead before writing a single line of code.

 

Part II: Architecting with Claude 


 

Start with Documentation, Not Code Before generating code, create a persistent foundation of documentation:


  • CLAUDE.md: Your project's persistent context (tech stack, standards, testing, security).
  • ARCHITECTURE.md: System components, data flow, and security boundaries.
  • FEATURES.md: Explicit feature definitions to prevent Claude from assuming a feature is "done" when it isn't.

Use Claude as Your Architect Your first prompt shouldn't be "Build the app." It should be:

"Read CLAUDE.md, ARCHITECTURE.md, and FEATURES.md. Act as a Principal Architect. Review the architecture. Identify design flaws, scalability concerns, and missing components. Do not write code."

This single step often saves days of rework. Always design database schemas and API contracts before implementation.

 

Part III: Implementation Planning 🗺️

The Plan-Act Loop Your mantra must be: Plan first, ship in slices. Let speed amplify good process—never replace it.

  1. Plan: Force Claude to think through the approach.
  2. Review: Check the plan across multiple axes (architecture, dependencies, scaling, single points of failure).
  3. Act: Once approved, let Claude execute.

Use Plan Mode for Complex Tasks Using Plan mode is like repeatedly aligning requirements with a PM before development. Once the plan is confirmed, execution can be handed over to the AI to run automatically with minimal supervision.

 

Part IV: Code Implementation at Scale 

Run Multiple Sessions in Parallel The biggest productivity unlock is running 3–5 Claude sessions in parallel, each in its own git worktree. Use claude --worktree in the CLI. This breaks the traditional linear development mindset.

Multi-Agent Orchestration For large projects, use a lead-agent model. Prompt Claude to create a team, spin up teammate agents in isolated worktrees, and assign tasks (e.g., Backend API, Frontend React, DB migrations).

Delegate Research to Sub-Agents Context dilution is real. Keep your main thread focused by delegating heavy research and file-reading to sub-agents.

Implement Institutional Memory Set up a 'memory' folder for agents to automatically store important findings. This creates institutional memory that persists across sessions, preventing agents from redoing trivial research.

 

Part V: Code and Technical Reviews 

 

Verification is the #1 Practice The Claude Code team emphasizes that the most impactful tip is verification—giving Claude a way to check its own output. Use automated tests, CI/CD pipelines, and review agents. If you only adopt one practice from this article, make it this one.

Don't Let Claude Touch Git Directly Allow Claude to do most things, but reserve Git privileges for yourself. Do frequent diffs, commits, and pushes. Comprehension ≠ compliance. In one documented case, an agent pushed directly to main despite explicit instructions to open a PR—and then pushed its own apology directly to main!

 

Part VI: Security and Production Readiness

  • Containerize Everything: Claude might attempt privilege escalation (like writing to /etc/shadow). Docker with no host mounts and restricted networks is non-negotiable.
  • Use Bot Accounts: Give each agent instance its own GitHub bot account for transparency and fine-grained permissions.
  • Make Tests Reliable: Flaky tests are worse than no tests. If CI is unreliable, the agent will use every failure as an excuse. Force the test harness to be reliable so it has to actually fix its bugs.
  • Define Hard Rules: Encode critical rules in CLAUDE.md (e.g., "Every function gets a test," "Never use global mutable state," "Insist on descriptive error messages").

 

Part VII: Team Standardization

Left to their own devices, developers will build fragmented Claude setups. A team with a shared configuration is 3-5× more effective than devs working individually.

Standardize your CLAUDE.md, permission configs, and security practices across the team. Regularly let Claude update CLAUDE.md itself—it's essentially reinforcement learning through agent feedback.

 

Conclusion: The Elevated Tech Lead

The technical lead's role in the age of AI is not diminished—it is elevated. You are no longer just writing code; you are:

  1. Framing the constraints and plans.
  2. Teaching agents how work should be done.
  3. Orchestrating parallel agents across worktrees.
  4. Reviewing plans and code rigorously.
  5. Standardizing team configurations.

Sometimes, it's more satisfying to watch the agents operate than to ship the feature yourself. But satisfaction aside, the real measure is impact. With the right practices, a single tech lead can orchestrate what once required entire teams.

Discipline is the only thing that makes AI coding reliable. Plan first, verify always, and never confuse speed with quality.

 

Over to you: What has been your biggest challenge in transitioning from writing code to orchestrating AI agents? Have you experimented with git worktrees or multi-agent setups yet? Let me know in the comments below! 👇

(If you found this valuable, consider subscribing to my Twitter or Plus91's AI Academy for more insights on engineering leadership and AI orchestration.)

 


Thursday, April 26, 2012

Providing Patients with Access to their Personal Medical Records On-The-Go

These are the slides from my talk on "Providing Patients with Access to their Personal Medical Records On-The-Go" at the 2nd Annual Wireless Healthcare Asia Summit 2012 in Singapore. Instead of talking about hard core tech and uber cool technologies, this time I touched upon the issues which have hurt the adoption of patient record technologies and driven the usage down, and caused consequent fragmentation of potential solutions. I strongly feel that actively driving adoption is the key to realizing the benefits of healthcare technology. And my talk was about a 3 point strategy to electronic patient records. INTEGRATE , ENGAGE AND SHARE...

Wednesday, February 09, 2011

NDNC just got a bit "Smarter"

The National Do Not Call (NDNC) registry is a short sighted solution to the nation's tele-marketing woes. Its strange that in a country built on a pedantic education system, the real life solutions which are implemented never seem to identify educating masses as a possible option. 

Note: to those who implemented the NDNC: your intentions were great; However, its important to look at the impact on Indian business of all types... The fact is that Mass SMS solutions are used not just for Tele Marketing. Since a Bulk SMS plan (primarily for telemarketers) is sold for 5-6 paise per SMS vs a Bulk Priority SMS plan for 50 paise or more per SMS (For Priority SMS which routes around the NDNC), Most Indian firms(primarily SME's) were happily using the former for all their SMS needs: even for critical alerts...Customers usually would demand that decision due to the knowledge that a lower price is possible and most firms would jump at this opportunity to inflate margins; Rashmi Bansal and many others celebrate this spirit and call it "Jugaad"...

Well the NDNC was appreciated at first, but as it went along, it started having detractors. The ones who were affected: All the way from Solution providers, eventually to the ones who wouldn't receive Password change alerts, Medical Appointment Alerts, and so on and so forth became critics of the NDNC. A most interesting and worth mentioning example here is of a "Software Security" Institute which would send the password for newly registered  users via an SMS and since half the nation had parked their numbers into the NDNC these guys would wait endlessly for the SMS ... I found it really silly on the part of this Institute: apparently teaching cutting edge IT Security,  they could'nt take the basic effort to update their systems: Instead they put up disclaimers stating that this is not their fault but that of the govt. :).

But what else does one expect. In India, entrepreneurs and  businessmen are driven by profit rather than anything else. Most Indian Business firms start talking RoI on day 1. The usual plan is to create quick and dirty solutions, as long as they work... and move to quick and dirty solution No. 2. Well getting back to the point, a number of cool firms locked in SMS solutions for their effectiveness and also for the coolness factor involved. The day NDNC came in, these guys were on plan 5 and plan 1 got hit. Most didn't care sadly.

Thankfully, the NDNC solution has been modified and made a bit more flexible and smarter. Now it allows mobile subscribers to register into the NDNC; but also specify their options on industry types do they find worth their time.

One simply has to dial 1909 from the mobile or send an SMS as following:

START [OPTION] and send it to 1909.

There are a number of options to choose from. For example, if you’d like to block all calls and messages, your option code is ‘0’ (ZERO). In this case, you should send an SMS as START 0 to 1909. Other options are:
  1. To block all calls and SMS except from Banking/ Insurance/ Financial products/ Credit cards category
  2. To block all calls and SMS except from Real estate category
  3. Block all calls and SMS except from Education category. Similarly,
  4. Allow only health related calls or SMS
  5. Allow only consumer goods including automobiles related calls or SMS
  6. Allow calls or SMS only from Communication/ broadcast/ Entertainment/ IT related fields
  7. Allow calls or SMS only from tourism and Leisure segments.
As an example, to allow calls and SMS's from the Health care Sector and to block everything else
Compose an SMS "START 4" and send to 1909

On an other front, I am happy that more people can benefit from Plus91 's SMS Based Appointment AlertsPrescription Reminders. Further now more patients can take advantage of our Information Therapy Initiatives such as My Fertile Time. 


Thursday, July 12, 2007

How Low is low cost

A Few Years ago, while working with a Firm which developed Mobile Phones, the Question always came up how much time before we create a sub Rs. 1000 handset. We already knew Nokia, Motorola and a lot of chinese vendors were working on the same, and firms like TI were developing chipsets for enabling phones. Today, i still dont find a phone like this, but its Definitely Impressive how the Low Cost Market Today is Loaded with Handsets of different Shapes, Sizes, Applications and Brandnames.. Here in India, where a neat phone with a Camera, GPRS and a few Cool Applictions with support for MP3 playback would'nt cost less than Rs. 10000, now we have a huge choice of such products in the sub Rs 5000 range. I recently tried out 2 of the 3 models launched by Reliance: The Classic Handsets(Made by ZTE), and was seriously impressed. I know they are CDMA and a lot less goes into designing these babies, they are developed possibly in China,, and definitely the overall cost of manufacturing is low, but still the consumer really does'nt care. What they see is some cute looking phones, priced around Rs. 1250 and offering plans which really don't burn a hole in the packet... The entry of Chinese vendors in the market is thus in a way a good sign, and I hope the Indian firms developing handsets catch on fast and start releasing their brands fast. There has been news of BPL, Hutch, Reliance, Elcoteq, Ajanta, Digibee, Spice and a whole load of other firms developing Cell phones in India. I'm sure companies like Digibee and Spice will be flooding the market in the near future...I just hope that its fast and its worth it...

// Nrip Nihalani