Senior AI Software Engineer

Role : Senior Software Engineer

For Lecxe : You design the system before you write the code, you build every week, and you do it inside our own proprietary AI engineering system.

Role : Senior Software Engineer

Team : Engineering

Location : Gurgaon (in-office, with client travel across India)

Reports to : Senior Product Manager

Stack : Next.js, FastAPI, Postgres, LangGraph

Type : Full-time

About Lecxe :

Lecxe is an AI-native consultancy building Business Operating Systems for Indian MSMEs. We solve real, unglamorous problems : agentic ERP adoption, inventory management, vendor negotiations and distribution, HR, finance, an AI personal assistant, and voice notes on WhatsApp turned into clean data. We ship software that fits how a business already works, instead of forcing it to change. We work directly with founders and operators, ship in weeks not quarters, and our work shows up in their daily business the day after we deploy.

About the Role :

Forget what you think a senior developer does. Here you design the system, then you build it. Does this need a new table, or not? Should this run on the server or in the browser? Should the filtering happen in the database or in the UI? You make those calls, and then you write the code. Nobody hands you a ticket with the answer already in it.

You also get to use Lecxe Compass, our own proprietary system for building reliably with AI. Custom agent harnesses where context is engineered and fed to the agents deliberately, not left to chance. Guardrails and evaluation loops we wrote ourselves. Most teams are still typing prompts into a chat window and hoping for the best. We are several steps past that, and it is why our engineers ship at a speed that does not look normal. You get to use Compass from your first week, and you get to make it sharper.

The stack is Next.js, FastAPI, Postgres and LangGraph. Tools you already know, on problems you have not seen before. You design the screens as well as the system behind them. You sit with the owner whose business you are changing. And you watch it go live in weeks, not quarters.

What You'll Do :

Architecture & System Design :

- Own the architecture on every project : the agents (LangGraph), the API (FastAPI), the database (Postgres) and the frontend (Next.js). You build phase one while keeping phase two and three in mind.

- Build every week, hands on the keyboard. Make the calls as you go : build or buy, new table or no new table, server or client, store it or compute it. The rest of the team builds on top of what you decide.

- Write data models and API contracts clean enough that a junior developer can extend them six months later without asking you a single question.

- Catch technical debt early and fix it with a plan, instead of at 2am with a hotfix.

AI-Native Engineering : Our Proprietary System :

- Work inside Lecxe Compass, our proprietary engineering system : custom agent harnesses, deliberate context management, and evaluation loops built in-house. This is the difference between Engineered AI that is reliable and Generic AI that is merely fast.

- Build and lead a team of AI-native engineers. Teach them when to trust an agent, when it is confidently wrong, and when it has written code that passes every test but misses the point.

- Ship real AI features : multi-step agent pipelines, WhatsApp voice notes turned into clean data, documents read and understood automatically, and search over a company's own files that works in production, not just in a demo.

- Push Compass forward. What you learn on client work feeds straight back into the harnesses and the context layer, so you are building the system, not just using it.

Design : Beautiful Products, Not Just Functional Ones :

- Own the visual and interaction design of what you build. Business owners and their staff use Lecxe products every morning, so they should feel considered, not generated. You know what good looks like, and you use it.

- Think in user flows before components. Before you open your IDE, you already know who's using this, what they're trying to do, and what friction you're removing.

- Set the design standard : component libraries, spacing systems, interaction patterns. A junior developer building a new screen should be able to follow your patterns without asking.

Engineering Standards & Team Leadership :

- Set and enforce engineering standards across the team : code review, automated testing, deployment hygiene, observability, and security baseline. You are the bar.

- Run code reviews that teach, not just that fix. Your review comments make junior engineers better, not just compliant.

- Onboard new engineers to ship within their first week. Orient them on the codebase, the patterns, the tools and the culture, then get out of their way.

- Build the internal culture of how this team engineers : how we use our own tooling, how we review AI-generated code, how we decide when to trust the agent and when to override it.

Communication : The Engineer Who Can Talk :

- Walk clients through what you built, why you built it that way, and what it means for their business. No jargon, no slides, no PM needed to translate.

- Write documentation that a non-engineer can read and a junior engineer can build from. Crisp, specific, and free of ambiguity.

- Communicate proactively when something is off-track : to the Senior PM, to the client, to your team. No one at Lecxe finds out about a problem from a client before they hear it from you.

What We're Looking For :

- 3 - 5 years of software engineering experience with at least one production system you owned end to end : architecture, implementation, deployment and maintenance.

- Strong working knowledge of TypeScript/Next.js, Python/FastAPI, and Postgres. You are comfortable across the full stack and you have opinions about how each layer should be built.

- Hands-on experience with AI coding agents (Claude Code, Cursor, or equivalent) as a daily engineering practice, not a novelty. You know how to get the best out of them, and you know their failure modes.

- Genuine design sensibility. You can look at a UI and immediately identify what is wrong with it. You have built interfaces that felt good to use, not just interfaces that worked.

- Experience with LLMs in production : prompt engineering, agent orchestration (LangGraph or equivalent), tool use, RAG pipelines, and evals. You have shipped an AI feature that real users depend on.

- Excellent spoken English. You can explain a technical architecture to a non-technical founder, run a go-live session with a business owner, and give a junior developer feedback that lands without bruising.

- High agency. You do not wait for a ticket. You see what needs to be built, you design it, you build it, and you tell people what you shipped.

- Comfort with ambiguity and fast-moving scope. MSME clients change their minds. Deployments break. You stay calm, diagnose fast, and fix it.

- The ability to hold a lot of context at once. Our clients span manufacturing, schools, restaurants, law firms and publishing. You will be carrying several unfamiliar businesses in your head at the same time, and you get useful in days, not weeks.

Even Better If :

- You have led or mentored a small engineering team, even informally. You know what it takes to make other engineers better.

- You have built something with a design you are genuinely proud of : a side project, an internal tool, a client product. Bring it to the interview.

- You have open-source contributions, a GitHub with real commits, or technical writing that shows how you think.

- You have worked in a consulting, startup, or client-delivery environment where you were building for someone else's business, not just your own product.

What Working Here Actually Looks Like :

You lead the innovation here. What you build does not already exist somewhere else for you to copy. You work with people who care about good design and good product, and who get excited by the same things you do. Your work changes how a real business runs : the owner uses it every day, and the business grows because of it. And you move fast. A decision you make at 10am can be live by 5pm. There is no middle management sitting between you and the work.

This is what next-generation engineering looks like from the inside. Not a future prediction. Not a conference talk. The actual thing, right now, building real products for real businesses.

How We'll Know It's Working :

- The products you ship look and feel better than anything the client has used before, and they say so without being asked.

- Junior engineers on your team are shipping production code faster and with more confidence than they expected, because of how you review and teach.

- Your architectural decisions hold up six months later. No one is cursing your schema in a post-mortem.

- You have made the team meaningfully faster using AI, and you have taught them to do the same.

- Clients know your name, trust your judgment, and ask for you on the next engagement.

Why This Role Is Interesting :

Most senior roles give you one thing : technical depth, team leadership, interesting problems, or a product that matters. This one gives you all four, plus the most interesting engineering environment in India right now. You will work with AI at a depth most companies are still only planning. You will build products that change how Indian businesses run. You will lead engineers who want to learn from you. And what you ship this month is in someone's hands next month. Bored, underused, or stuck in a process that slows you down? This is the exit ramp.

LaneOne Ventures Private Limited (Lecxe) is an equal opportunity employer.

Skills

System Design, System Architecture, FastAPI, PostgreSQL, Next.js, Artificial Intelligence, Python, TypeScript, LLM, RAG, Full Stack