Founding Engineer

Uplers; India; On-site; INR 300,000 - 600,000; Full time

Job summary

Uplers is seeking a senior full-stack engineer to own the platform build for a UK-based PE Report AI project. You will shape architecture, ship features end-to-end across Python engine, Next.js app, and Excel add-in.

This hands-on role requires ownership, fast delivery, and rigorous verification of financial outputs, with AI-assisted development and a focus on production-quality systems.

Qualifications

• Senior full-stack engineer with 5+ years experience.
• Proficient in Python and TypeScript in production.
• Experience with FastAPI, Next.js, React, and cloud providers.
• Experience with AI tooling and evaluation pipelines.
• Strong PostgreSQL and Supabase knowledge.

Responsibilities

• Own platform architecture and end-to-end feature delivery.
• Collaborate with founder, shipping daily and owning code from backend to frontend.
• Work with AI tools to build, verify, and harden outputs.

Skills

Python

TypeScript

FastAPI

React

Next.js

Office.js

openpyxl

Aspose.Cells

PostgreSQL

Supabase

LangGraph

Distributed workloads

Async programming

Cloud & DevOps

LLM AI

Tools

Azure/GCP/AWS

Job description

Experience: 5.00 + years
Salary: Confidential (based on experience)
Expected Notice Period: 7 Days
Shift: (GMT+05:30) Asia/Kolkata (IST)
Opportunity Type: Remote
Placement Type: Full Time Contract for 3 Months(40 hrs a week/160 hrs a month)
(*Note: This is a requirement for one of Uplers' client - PE Report AI Company from UK)

What do you need for this opportunity?

Must have skills required:

Aspose.Cells, ISO 27001, LangGraph, office.js, openpyxl, SOC 2, 0?1 Product Ownership, Async & Distributed Workloads, Cloud & DevOps, LLM/AI Product Engineering, PostgreSQL + Supabase, Python + FastAPI, React/Next.js, TypeScript

PE Report AI Company from UK is Looking for:

The role

• You will work side-by-side with the founder as the interim owner of the platform build: shaping the architecture, shipping daily, and carrying features end to end — Python engine, Next.js app, Excel add-in, and the infrastructure underneath all three.
• This is a hands-on, high-ownership role for someone who is comfortable being the only engineer in the room: you will make pragmatic calls, hold a high bar where it matters (correctness of financial numbers is the product), and move fast everywhere else.
• Development here is AI-accelerated — you should be fluent working with AI coding tools and directing them well, and equally capable of reading, verifying, and hardening what they produce.

What You Will Build (3-month Arc)

Month 1 — production foundations

• Take the platform from local dev to deployed staging + production: containerise the FastAPI parsing/population service and the Next.js app, CI/CD with the existing test and eval suites as gates, secrets management, backups, monitoring and alerting.
• Authentication, organisations and multi-tenancy on Supabase/Postgres (RLS), so sponsors and portfolio companies each see exactly their own templates, data and questions.
• Harden the Excel add-in for real users: hosted task pane, sign-in from the pane, and centralised deployment via the Microsoft 365 admin centre.

Month 2 — scale the engine''''s operations

• Re-architect long-running work (template onboarding, populate runs) onto a proper queue/worker model with streamed progress, retries, resumability and per-run cost caps — replacing the current in-process execution.
• Wire the evaluation suite into CI as a regression gate: every engine change is scored against the ground-truth labelled workbooks before it merges track fill accuracy, incorrect-write rate, LLM cost and latency budgets.
• Ship product surface with the founder: the review-question inbox, the template contract UX, onboarding visualisation, and the add-in fill experience.

Month 3 — pilot-ready.

• Support first sponsor pilots end to end: onboard their real templates, run their real data packs, close the loop on failures fast.
• Security posture for financial data: access controls, audit logging, data retention, and the groundwork for SOC 2.
• Reliability hardening, load and failure testing, and clean handover documentation for the permanent team.

Technical Requirements (must-have)

• Senior full-stack engineer (typically 5+ years) fluent in BOTH modern Python and TypeScript.
• Python: FastAPI, Pydantic, async patterns, and production service design comfortable in a large, test-covered codebase.
• TypeScript/React: Next.js App Router in production, component architecture, Tailwind.
• LLM product engineering: you have shipped features built on Anthropic/OpenAI APIs — prompt and context design, structured outputs, evaluation harnesses, cost control, and observability (LangSmith or equivalent). You know why ''''''''''''''''''plausible but wrong'''''''''''''''' is the failure mode that matters.
• PostgreSQL depth: schema design, migrations, row-level security Supabase experience is a strong plus.
• Infrastructure ownership end to end: Docker, CI/CD (GitHub Actions or similar), and at least one ofAWS/GCP/Azure/Fly/Render taken to production, including monitoring and incident response.
• Async/long-running workloads: queues, workers, streaming progress (SSE/WebSockets), idempotency and retries.
• Deep comfort with Excel as a data structure — formulas, references, formats, named ranges — and willingness to work at that level of detail daily.
• Evidence of 0→1 ownership: you have built and operated a real product with a very small team, and can show it.

Nice to have

• Office.js / Excel add-in development, or Aspose.Cells / openpyxl-class spreadsheet tooling.
• Private-equity or finance-adjacent domain: management accounts, flash reporting, KPI packs, consolidation.
• Agent orchestration frameworks (LangGraph or similar) and multi-model routing.
• Design sensibility — the product aims at a Harvey/Legora-grade bar.
• Security/compliance groundwork for financial data (SOC 2, ISO 27001).

How We Work

• Directly with the founder, daily — short feedback loops, decisions made in hours, not weeks.
• AI-accelerated development as the default human judgment on architecture, correctness and taste as the differentiator.
• Correctness is sacred: the eval suite and the deterministic verification layer are the product'''''''''''''''''s spine.
• Everything else is pragmatically fast.
• You will see the whole business, not a ticket queue: customer conversations, pilot feedback, roadmap trade-offs.