Artificial Intelligence Engineer (Web Agent)

Position Overview

We are looking for an

AI Engineer

with a minimum of 2 years of experience, preferably in the field of AI. We expect you to be hands-on working with various

LLMs (large language

models) and foundational AI models . This role is application-focused, requiring the ability to test, fine-tune, and deploy AI agents tailored for specific tasks. The ideal candidate should be well-versed in

open-source and proprietary AI solutions , selecting the most effective model for each problem.

You will work closely with our product and engineering teams to develop

scalable AI-driven applications , leveraging LLMs in real-world use cases. A strong understanding of model evaluation techniques such as

LLM as a Judge , fine-tuning methodologies, and prompt engineering strategies is essential.

Key Responsibilities

Work with different

LLMs (open-source & proprietary) , including OpenAI GPT, Anthropic Claude, Mistral, Llama, and open-weight alternatives like Falcon, Bloom, and others.

Evaluate and compare AI models to determine the best solution for specific applications.

Fine-tune and optimise AI models for domain-specific use cases, ensuring efficiency and accuracy.

Develop and implement

AI agents

capable of performing specialised tasks.

Define the best

training document formats and datasets

required for model fine-tuning and reinforcement learning.

Utilise

LLM as a Judge

framework to assess AI-generated responses and improve model accuracy.

Model Monitoring :

Tracks performance metrics like accuracy, latency, and drift to detect issues early.

Data Quality Checks :

Ensures input data remains consistent and relevant, preventing model degradation.

Bias and Fairness Audits :

Identifies and mitigates biases in AI predictions.

Explainability & Debugging :

Provides insights into model decision-making, making AI more interpretable.

Real-time Anomaly Detection :

Helps detect outliers or failures in production.

Automated Alerts & Logging :

Sends alerts when performance degrades, aiding in proactive issue resolution.

Deploy models via

APIs, cloud platforms (AWS / GCP / Azure), or on-prem solutions .

Collaborate with software engineers to integrate AI models into existing applications.

Stay current with the latest advancements in

AI model architectures, fine-tuning strategies, and inference optimisation .

Implement responsible AI practices, ensuring transparency, bias reduction, and ethical considerations in model deployment.

Required Skills & Experience

Proficiency with LLMs & AI Models : Hands-on experience with GPT-4, Claude, Llama, Mistral, and other foundational models.

Fine-tuning & Customization : Knowledge of supervised fine-tuning, reinforcement learning from human feedback (RLHF), and parameter-efficient fine-tuning techniques (LoRA, QLoRA, PEFT, etc.).

AI Model Evaluation : Experience using LLM as a Judge, OpenAI Evals, and other benchmarking tools.

Prompt Engineering & Retrieval-Augmented Generation (RAG) : Expertise in designing optimal prompts and using vector databases like FAISS, ChromaDB, or Pinecone.

Programming & ML Frameworks : Strong skills in Python with experience in

PyTorch, TensorFlow, JAX, or Hugging Face Transformers .

LLM Deployment & Optimization : Familiarity with tools like vLLM, TGI, DeepSpeed, Triton Inference Server, and LangChain.

Data Preparation & Processing : Knowledge of

tokenisation, dataset curation (HF Datasets, OpenWebText), and synthetic data generation .

API Integration : Experience working with OpenAI API, Anthropic API, Together.ai, Fireworks AI, or similar services.

Cloud & MLOps : Hands-on experience with

AWS Sagemaker, GCP Vertex AI, Azure ML, Kubernetes, Docker, and CI / CD for AI workflows .

One or more AI Tools and Frameworks for Observability - OpenAI Evals, Arize AI, WhyLabs, Fiddler AI, Evidently AI, Weights & Biases, Neptune.ai—a metadata store for tracking ML experiments and models, Prometheus & Grafana, MLflow – A popular tool for monitoring ML experiments, models, and deployments, Seldon Core

Version Control & Experiment Tracking : Experience with

Weights & Biases (W&B), MLflow, or similar platforms .

Nice-to-Have Skills

Experience working with

multi-modal models (text, image, video, audio) .

Knowledge of

autonomous AI agents

like AutoGPT, BabyAGI, or CrewAI.

Familiarity with

Graph-based AI models

and knowledge graphs.

Understanding of

privacy-preserving AI

and federated learning.

Hands-on experience in

low-latency inference optimisations

(FlashAttention, quantisation, distillation, etc.).

Why Join Us?

Work on cutting-edge AI applications with real-world impact.

A dynamic, fast-paced environment with

end-to-end ownership

of AI-driven projects.

Competitive salary, benefits, and opportunities for career growth.

A culture of learning, innovation, and collaboration.

Perks

Compensation that rewards your mastery, supplemented with performance-driven incentives.

A wholesome package of training and developmental avenues that constantly enrich your skill set. An ecosystem fostering innovation, where every voice harmonizes into the choir of

progress. A chance to script your chapter in Nbyula's success saga celebrated with fervor.

Who is an ideal match for being a terraformer at Nbyula? All the attributes that we are looking for in an ideal teammate.

Openness- We welcome people from different backgrounds and schools of thought. Terraformers are open to different perspectives in approaching a solution and do not just limit their thoughts or ideas to only a specific domain.

Conscientiousness- We believe in working together for the larger goal and with complete dedication and not just for personal benefits, however, we do not expect terraformers to work to the point of burnout.

Humility- Being humble, grateful and respectful are the core traits of terraformers, we do not expect people to agree with every view of the management, feel free to have a different perspective, but we always expect it to be put forward with respect.

Risk Takers- Terraformers are not afraid of the unknown and are open to new things, not that we encourage extreme risks without weighing the consequences but we take calculated risks.

Autodidacts- Terraformers teach themselves to learn, we do our research to get solutions, and we do not expect you to have a blank slate and figure everything out yourself, we are here to guide you but not handhold and micromanage you Self-Actualization- Terraformers are on the path of self-actualization, we are not bothered by the noise and distractions around us, we only work towards achieving our full potential. We do not expect you to over-burden yourself and not have fun, but we expect you to work to the best of your capabilities

About Us : Nbyula is a German technology brand headquartered in Berlin with a Research & Development Center in Bengaluru, Karnataka, operational since 2014. Nbyula believes in creating an open world where opportunities and talent are not hindered by borders or bureaucracies. Nbyula is materializing this by leveraging the bleeding edge of technologies like cloud, distributed computing, crowdsourcing, automation, gamification, and many more. The North Star is to create a horizontal marketplace encompassing people, content, products & services for international work and studies to enable, train and empower "Skillizens without Borders''.

To learn more about us, please visit https : / / nbyula.com / about-us