Top-of-market pay + ESOP
Benchmarked twice a year. Equity for every full-time role.
Life at ChainCraft
Built by builders, for builders
You'll work in a team of 3–5 seniors with direct client access and real decision power. We write things down, demo weekly, and retro honestly. Most of our leads started as engineers here — we promote from within because we hire people worth promoting.
3–5
people per project team
9+
yrs avg. engineering experience
0
hours on a bench
4.8
Glassdoor rating
Benefits
Taken care of, properly
Remote-flexible
Work from anywhere in India; quarterly team weeks in Mumbai.
Learning budget
₹1.5L/year for courses, conferences, and compute for side projects.
Health, fully covered
Premium insurance for you, your family, and your parents.
Real time off
24 days leave plus company-wide winter shutdown. Actually taken.
Latest hardware & models
Top-spec machines and unmetered access to frontier AI models.
What the work is actually like
You will be on one project at a time, in a team of three to five, with direct access to the client’s engineers and decision-makers. There is no account manager between you and the person whose problem you are solving, and there is no separate team that scoped the work before you arrived — you were probably in the room.
A typical week has one live demo on real client data, a written update you contribute to, and four days of building. We default to writing: decisions, architecture, trade-offs and post-mortems all get written down, partly because clients need the record and partly because writing is how we find out whether we actually understand something.
The work is unusually varied for a delivery role. Over a year you might build an agent that handles logistics exceptions, a retrieval pipeline over ten years of clinical guidelines, and a voice system that books appointments for forty dental practices. The domain changes; the engineering standard does not.
It is also unusually accountable. Every system we ship launches with agreed KPIs and a dashboard. You will find out whether your work moved the number, which is uncomfortable occasionally and is the main reason people who join tend to stay.
How we are structured, and what that means for you
There is no bench and there are no utilisation targets. Between projects you are working on internal tooling, the evaluation frameworks, or the research that feeds our writing — not being marketed to a client as available capacity.
We promote from within because we hire people worth promoting. Most of our leads started as engineers here. Career progression is discussed openly twice a year alongside compensation, which is benchmarked at the same cadence rather than negotiated when someone threatens to leave.
Seniority at ChainCraft means judgement rather than tenure: knowing which problems are worth solving, when to tell a client no, and when the correct engineering answer is something simpler than the interesting one. We interview for that specifically, which is why the technical round is a conversation about systems you have built rather than a puzzle.
Open positions
We're hiring
Design and ship agent systems and retrieval pipelines that run live client operations — including the guardrails, evaluation harnesses and observability that make them trustworthy.
What you'll do
- Architect agent and RAG systems against real client workflows
- Build evaluation suites and drive measured quality improvements
- Own guardrail design: action whitelists, approval gates, injection defence
- Work directly with client engineers and demo weekly
What we look for
- Five or more years building production software, with recent LLM system experience
- Strong Python or TypeScript, and comfort with distributed systems
- Evidence of shipping something that stayed in production and was maintained
- The judgement to tell a client when an idea will not work
Build AI-native web products end to end: streaming interfaces, multi-tenant backends, billing and metering, and the interaction patterns that make probabilistic output feel controllable.
What you'll do
- Ship production Next.js and TypeScript applications with streaming AI interfaces
- Design multi-tenant data models with isolation enforced at the database layer
- Implement metering, billing and cost attribution for AI features
- Own performance and Core Web Vitals on the products you build
What we look for
- Four or more years of full-stack work, deep in React and TypeScript
- Experience with Postgres and a real understanding of tenancy and access control
- An eye for interface detail — our design bar is high and non-negotiable
- Interest in the UX problems specific to probabilistic systems
Own the infrastructure every client system runs on: orchestration, model gateways, evaluation pipelines, observability and cost control.
What you'll do
- Build and operate orchestration and model-gateway infrastructure
- Instrument cost, latency and quality across every deployed system
- Automate evaluation runs in CI and gate promotions on them
- Deploy into client VPCs across AWS, Azure and GCP
What we look for
- Four or more years in infrastructure or platform engineering
- Kubernetes, Terraform and at least one major cloud in depth
- Experience operating systems where an outage has a business consequence
- Pragmatism about tooling — we choose boring technology deliberately
Design interfaces for systems that are fast, useful and sometimes wrong — streaming, citations, confidence, review and undo — across client products and our own.
What you'll do
- Design AI interaction patterns and the states most design systems lack
- Prototype against live models, because these interactions cannot be judged statically
- Own design systems that client engineering teams will extend after we leave
- Work directly with users during discovery
What we look for
- Four or more years designing software products, with a portfolio that shows craft
- Comfort working in code-adjacent tooling and reviewing implementation
- Genuine interest in designing for uncertainty rather than hiding it
- Strong written communication — we default to writing
Run discovery engagements that produce ranked, costed, buildable roadmaps — and the governance frameworks that let organisations act on them.
What you'll do
- Lead opportunity discovery and feasibility scoring across client functions
- Build cost and ROI models grounded in what our engineers actually charge to build
- Design AI governance frameworks and regulatory exposure maps
- Present to executive teams and to the engineers who will implement
What we look for
- Five or more years in consulting, strategy or product, with real technical literacy
- Ability to inspect data and judge feasibility rather than assuming it
- Willingness to recommend against work, including our own
- Excellent written output — our deliverable is a document people act on
Hiring process
Four steps, two weeks, no puzzles
- 01
Intro call (30 min)
Mutual fit, your story, our honest pitch. With the hiring manager, not a recruiter.
- 02
Technical deep-dive (60 min)
A conversation about systems you have built — no leetcode, no trick questions.
- 03
Paid work sample (½ day)
A realistic, scoped problem from our actual work. Paid, and yours to keep.
- 04
Team fit + offer
Meet your future team, ask anything. Offer within 48 hours of the final round.