Job Description

About Applied Computing

Applied Computing builds foundation models for the most valuable and consequential sectors on the planet, all within the energy chain. We came out of stealth less than a year ago and have already seen exceptional success. We're now looking to grow our team with world-class minds who are driven to grow into the best version of themselves and to put their shoulder behind the wheel of applied AI, so we're recruiting an

AI Principal Research Scientist.

The Role

We're looking for a Principal Research Scientist to spearhead the design and productionisation of our core AI backbone--an intelligent query router, a high-throughput retrieval-augmented generation pipeline, and a physics-informed time-series forecasting engine.




In this role, you will drive end-to-end innovation: crafting novel multi-agent orchestration strategies, inventing hybrid statistical/deep models with built-in physical constraints, and hardening research prototypes into containerised microservices that meet strict latency, scalability, and reliability targets in AWS.

Location:

We have a team distributed across US, Europe, India & Latam. Ideally you are based in a convenient time zone for collaboration so East Coast US, Canada, Europe, India etc.

What You'll Do

Drive Research & Architecture

Invent and evaluate novel multi-agent orchestration strategies, defining the API and decision logic that routes queries between language models, knowledge-base lookups, and analytic services Lead the development of advanced RAG techniques: hybrid retrieval (e.g. dense + sparse), neural reranking, chain-of-thought summarization, and end-to-end Q&A evaluation frameworks.

Prototype & Productionise

Translate research prototypes into production-grade Python modules, leveraging ML frameworks (PyTorch preferred), containerisation (Docker, Kubernetes/ECS/EKS), and AWS services (SageMaker, Lambda, S3, RDS). Design and build CI/CD pipelines that automate data ingestion, model training, validation, canary deployments, and continuous retraining triggers.

Benchmark, Validate & Iterate

Establish rigorous evaluation protocols and back-testing suites for each subsystem (e.g. precision/recall for RAG, calibration metrics for forecasting intervals, routing accuracy and throughput for the orchestrator).

Mentorship & Collaboration

Guide researchers and ML engineers through code reviews, whiteboard sessions, and design critique.

Architect Multi-Agent Orchestration

Design and implement dynamic routing policies that dispatch incoming user queries to the most appropriate AI agent or service (LLM, vector search, analytics engine) based on intent, cost, latency, and confidence. Build a pluggable plugin/adapter framework so new agents (e.g. specialised LLM chains, external APIs) can be onboarded with minimal code changes.

Dynamic Decision Logic & Fallbacks

Develop sophisticated decision trees and fallback strategies (e.g. retry with a smaller model, degrade to a cached answer, escalate to a human-in-the-loop) to guarantee both reliability and cost-effectiveness. Implement health?check and circuit-breaker patterns so that any under?performing agent is automatically sidelined.

Must Have

PhD in Computer Science, Statistics, Applied Mathematics, Physics, or a closely related field. 5+ years of applied research experience bridging ML theory and production systems. Publication record as first or senior author in at least two of: multi agent systems, vision-language modelling, computer vision, information retrieval and/or large-scale LLM systems. Deep expertise in Python and ML frameworks (PyTorch preferred), with proven ability to write maintainable, testable, and container-ready code. Hands-on experience with LLMs and retrieval-augmented generation: vector search (e.g. FAISS, Annoy), BM25, neural rerankers, agentic stack. Production skillset: Docker, Kubernetes or AWS ECS/EKS, CI/CD (GitHub Actions, CodePipeline), AWS ML services (SageMaker, Batch). Data engineering acumen: ETL pipelines, relational and time-series databases (Postgres/TimescaleDB), search/index stores. Advanced Python Skills: Deep knowledge of asynchronous programming (asyncio, FastAPI, or similar), gRPC or HTTP/2, connection pooling, and performance tuning. Plugin & Adapter Frameworks: Hands-on experience building pluggable architectures or SDKs that allow new "agents" or downstream services to be onboarded with minimal friction. Excellent communication: Able to author research-style documentation, present to executives, and translate complex ideas for cross-functional teams.

An Insight into Culture

We don't just hire for skills -- we hire for trajectory.

Whether you're an AI researcher, machine learning engineer, software engineer, GTM strategist or operations builder, we're looking for people who raise the bar. Not just technically, but in how you communicate, lead, and execute. If you've ever felt underused, under-challenged, or stuck in slow-moving / bogged down teams and projects -- we are your reset button.



We write our own rules.


We care more about your ambitions than your CV.
We hire engineers who ship, AI researchers who bring real world impact, and commercial talent who think like entrepreneurs.




Finally, if you are highly autonomous, low-ego, and aligned with Applied Computing's mission come talk to us.