What You'll Actually Do

Model Development & Deployment

• Design, build, and deploy ML models end-to-end, from problem framing through to production serving and performance monitoring
• Know when a model architecture choice, training approach, or data assumption is likely to become a quality or delivery risk, and fix it early
• Build and optimise NLP systems: Q&A, sentiment analysis, entity extraction, intent classification, and semantic search
• Develop and ship LLM-powered applications across a range of industries and use cases

Data & Pipeline Ownership

• Design and own the data pipelines that feed your models vs. just consuming what a data engineer hands you
• Build document comprehension and information extraction pipelines using deep learning and OCR
• Create RAG-based systems that dynamically build knowledge bases from unstructured data

Collaboration & Technical Leadership

• Work directly with enterprise clients to translate business problems into ML approaches
• Engage meaningfully with architects, engineers, and product stakeholders; your technical judgement will carry weight
• Contribute to Ideas2IT's AI practice: internal tools, frameworks, and the standards that shape how we build

‍

What You'll Bring

Must-Haves

• 3–8 years of hands-on Data Science experience: statistics, ML, and Deep Learning
• Strong Python skills and software development fundamentals
• Strong grasp of Transformer architectures and modern LLM fundamentals 
• Foundational understanding of classical deep learning architectures (RNN, LSTM, CNN); you know why they existed and when they still apply
• Production experience with at least one DL framework: PyTorch, TensorFlow, or Keras
• NLP experience: NLI, Q&A systems, sentiment analysis, entity extraction
• Experience with semantic and syntactic pre-processing pipelines

Good-to-Haves

• Hands-on exposure to LLMs, Prompt Engineering, LangChain, or RAG pipelines
• Agentic AI development using cloud or open source frameworks

Why Ideas2IT, and Why Now?

Most data science roles at services firms are maintenance plays: tuning models built for yesterday's problem, on pipelines someone else owns. Ideas2IT has operated AI-first from the ground up: proprietary platforms in production, agentic tooling embedded into delivery, 800+ engineers already AI-fluent. You won't be making the case for doing real AI work. You'll be doing it.

We've built and shipped platforms including Legacyleap (GenAI-powered legacy modernisation), DataStoryHub (AI-powered BI with natural language querying), and Factory.360 (industrial IoT). These are not prototypes or pilot projects, but are live, generating revenue, with actual customers. The models and systems you build here will be the intelligence layer underneath all of it.

‍

What that looks like in practice:

• You'll work on projects with real stakes: AI platforms for clinical trials, healthcare data infrastructure, financial services, and more
• You'll shape the approach you execute, with your architectural and modelling decisions 
• You'll pitch ideas at internal Shark Tank-style sessions and Hackathons: if you want to build something, there is a real path to doing it
• Our founder's guarantee: you'll always be part of at least one initiative that genuinely interests and challenges you

‍

The Hiring Process: We Mean Business and Move Fast

• Send us something real: your resume, a short note on the models and systems you've owned. We're not fussy.
• Do a real-world challenge: we'll share a scenario relevant to the work you'll do in this role.
• Talk to the people you'd work with: connect with the people who'll be your colleagues.
• Get a quick, respectful answer: yes or no, with a reason either way.

‍