Skip to content
All work
LLM Application2024

JobAssist AI

An AI job-search platform for the French market — CV building, cover letters, application emails, and role matching.

A job-search platform aimed at applicants in France, built around four AI surfaces: a CV builder, a cover-letter generator, an application-email generator, and AI-driven job matching and recommendations. Each surface has to produce output a candidate can actually send without rewriting it.

Primary technologies

  • Python
  • LLM integration
  • Prompt engineering

01 / Context

What this project is

A job-search platform aimed at applicants in France, built around four AI surfaces: a CV builder, a cover-letter generator, an application-email generator, and AI-driven job matching and recommendations. Each surface has to produce output a candidate can actually send without rewriting it.

Engineering problem

Generated application material fails the moment it reads as generic — it has to reflect the specific candidate and the specific posting, in the conventions the French hiring market expects.

Approach

Built the AI layer as a set of focused generators — CV, cover letter, application email — alongside a matching and recommendation system, rather than one general-purpose prompt doing everything.

02 / Architecture

How the system is structured

  1. 01

    Layer 01

    Focused generators

    Separate CV, cover-letter, and application-email generators rather than one general-purpose prompt serving every surface.

  2. 02

    Layer 02

    Matching & recommendation

    Ranks openings against a candidate's profile to surface roles worth applying to.

03 / Engineering judgment

Decisions that shaped it

  1. 01

    One generator per artefact

    A CV, a cover letter, and an outreach email have different conventions and constraints. Splitting them made each output usable without a rewrite.

04 / Stack

What it's built with

AI

  • Python
  • LLM integration
  • Prompt engineering

05 / Capabilities

What it does

  1. 01

    AI CV builder

  2. 02

    AI cover-letter generator

  3. 03

    AI application-email generator

  4. 04

    AI job matching and recommendations

06 / Challenges

What was hard, and what fixed it

Challenge 01

Problem

Generated application material fails the moment it reads as generic — it has to reflect the specific candidate, the specific posting, and French hiring conventions.

Solution

Built each artefact as its own generator with its own constraints, instead of one prompt trying to cover every case.

Next step

Have something like this
you need built properly?

I work on the AI layer — retrieval, reasoning, and the plumbing underneath — and hand it over ready to integrate.

© 2026 Md. Sakibur Rahman