Ho presentato la mia candidatura tramite un selezionatore. La procedura ha richiesto 6 settimane. Ho sostenuto un colloquio presso Cactus Communications (Hyderabad) nel mese di apr 2022
Colloquio
it has 4 rounds of interviews.
1. Initial technical round ( basics of ML, resume based too)
2. Take-home challenge
3. depth ML round ( Ml/Dl algorithms, in-depth NLP, transformers, etc.)
4. Hiring manager round ( resume/role-based, behavioral questions
Domande di colloquio [1]
Domanda 1
Explain Logistic regression, Explain positional encoding, how to train BERT from scratch ..
Ho presentato la mia candidatura online. La procedura ha richiesto una settimana. Ho sostenuto un colloquio presso Cactus Communications (Abu Dhabi, ) nel mese di gen 2022
Colloquio
This company is a complete waste of time. And their Cactus Lab which they formed after acquiring another company in Denmark has some legit people who knows about AI, rest if you happen to talk to someone from India at an executive position then they lacks complete knowledge around things and will only result in rejection of you demand a justified raise.
I had completed all the rounds which includes 2 technical, 1 techno management and HR discussion plus a pathetic questionnaire.
Post completion of each round the TA spoc said that things are positive and I should hear within a day or 2 and "They are looking for you to work for Language Model training" and soon he shall discuss the offer. Next day he called me and said that my skillsets are not matching with what they are looking for. When further probed, he said that feedback is since I've never created any custom neural network, I don’t qualify for their so called position. I clearly don’t understand was the panel sleeping during the interview. And are the interview panel so incompetent that they just couldn’t see if I don't fit the role in the first 2 rounds.
From where I see it, this is just a budget issue. And reading company reviews here just supports this fact.
Overall I would suggest that please don't fall for this company, it is cheap low paying and a waste of time.
**Another highlight which I found during technical rounds is their AI products are still at very early stages and all editing is done manually, which they claim is done by AI.
Domande di colloquio [1]
Domanda 1
1) Explain all the NLP projects which I've done so far.
2) Described their NLP task and asked me to think of all possibilities around it.
3) How would you train a supervised problem with unlabelled dataset.