I applied through a recruiter. I interviewed at Axya in Apr 2024
Interview
Interview questions requiring recorded video answers, horrible experience as interviewers never got back to me, with any response. Monitored the position afterwards and doesn’t seem like anyone actually filled the position. I suspect this company is only using this information for free ideas.
I applied online. The process took 1 week. I interviewed at Axya (Montreal, QC) in May 2025
Interview
Interview process concerns: Second round required presentation + diagrams + prototype - excessive time commitment for candidates. Homework assignments appear related to actual company problems, which feels like unpaid consulting work. There are more ethical ways to evaluate technical skills without exploiting candidate time.
Interview questions [2]
Question 1
Axya has built an industrial procurement platform with many diverse types of customers. Each of
these customers have also a diverse pool of supplier who all have their unique ways of sending
quotations or other kind of procurement information in PDFs.
Each supplier’s follows a stable format per supplier but varies across suppliers.
Challenge:
1. Automatically extract structured quote fields (part numbers, unit prices, quantities,
delivery dates, payment terms) from heterogeneous PDF documents.
2. Provide a queryable service endpoint that returns normalized quotes in JSON.
Key Requirements:
● OCR & Layout Analysis: Propose OCR engines (e.g., Amazon Textract, Tesseract,
LayoutLM) and strategies to detect table/grid structures.
● LLM Integration: Outline how you would use a pre-trained LLM (or fine-tune) to correct,
normalize, and validate extracted text and map to schema.
● Scalability & Fault Tolerance: Design for high throughput and intermittent failures using
AWS primitives.
● MLOps Pipeline: Define CI/CD for pipeline updates, model versioning, automated
testing, and performance monitoring (e.g., SageMaker Pipelines, CloudWatch).
● Deliverable Service: A RESTful API or microservice specification that ingests a PDF
URL (or S3 URI) and returns a JSON payload of extracted fields.
The platform has thousands of aerospace suppliers with structured attributes (capacities,
certifications) and unstructured documents attached to them (HTML pages, PDFs). All of this
information has some commonalities, but a lot fo what makes each of these companies
successfully doesn’t necessarily fit a common schema.
A buyer for an aerospace company should be able to communicate a need in plain language
and receive a list of suppliers that match its requirements and the context surrounding the
request.
Note: The current system uses full-text ElasticSearch, and you can test it out here:
https://axya.co/suppliers_directory?page=0
Challenge:
1. Index structured and unstructured data into a unified semantic search solution to answer
capability queries (e.g., "CNC machining for titanium aerospace parts").
2. Make sure that part of the query that is deterministic gets treated as such (i.e. specific
certification required or geolocalisation of the suppliers).
Key Requirements:
● Data Ingestion & Preprocessing: Describe ETL for structured tables and document
parsing (PDF, HTML), metadata extraction, and cleaning.
● Embedding & Vector Store: Choose embedding models (e.g., OpenAI embeddings,
Sentence Transformers) and vector database architecture.
● “RAG” Pipeline: Illustrate how a retrieval layer and LLM can be combined to answer
free‐text queries with structured output (e.g., top-N supplier list with relevancy scores).
● Cloud Deployment: Architect an AWS-based solution for indexing, query API, and
autoscaling.
● MLOps & Monitoring: Propose a CI/CD process for retraining embeddings (if needed),
refreshing indexes, and tracking query performance and drift.
Note 1: Whenever possible, we much prefer to reuse existing technologies than to add new
ones.
Note 2: all of the information collected and used for indexing are public information from
suppliers.
Deliverables
1. Slide Deck: 12–15 slides covering both projects end-to-end.
2. Architecture Diagrams: Detailed AWS diagrams for each system’s components, data
flows, and failover strategies.
3. Code Snippets / Pseudocode: Examples of key modules (e.g., data ingestion, model
inference, CI pipeline definitions).
4. Security & Compliance Notes: Brief discussion on data privacy and access controls
(when necessary).
5. (bonus) Optional Prototype: If time permits, a minimal proof‐of‐concept (e.g., Jupyter
notebook or small Lambda function).
I applied online. The process took 3 weeks. I interviewed at Axya in Oct 2024
Interview
Entirely unprofessional and disorganized from the get go. First panel interview, one member was off camera on another call for 90% of the total interview, the next team member turned his camera off and went on mute until he had to leave 30 minutes in. They still wanted to proceed with my candidacy and in the follow up interview with what was supposed to be two leadership members, was only one and he was 10-15 minutes late without warning because he was in the car. They then provided zero feedback or even attempted to let me know they were moving on.
In general, it seems less start-up and more school project. The blatant disregard for basic professionalism and even a belief in your actual product would go a long way in finding and retaining talent.
HR was the only individual with any sort of professionalism and was responsive.
Thank you for your feedback, we are sorry that you had this experience with us during your interview process. We appreciate your time and will definitely work on this.
Thank you for applying at Axya and going through the process.
