TixelJobs
P
pulsecrafts solutions pvt ltdvia muse

AI Solutions Architect

REMOTEPosted 1d ago
ML EngineerLeadFull-time#remote

About the Role

Job DescriptionAI Solutions Architect – LLM Fine\-Tuning \& Graph\-Based AI

**Location:** Open to discussion
**Designation:** AI Solutions Architect
**Experience:** 7\+ Years (with strong hands\-on experience in AI/ML and Generative AI)

Role Overview

We are looking for an experienced **AI Solutions Architect** with strong expertise in Large Language Models (LLMs), model fine\-tuning, and advanced AI architectures. The ideal candidate will have hands\-on experience designing, developing, and deploying customized AI solutions using fine\-tuning techniques, Knowledge Graphs, and GraphRAG.

The candidate will be responsible for architecting scalable AI solutions, optimizing LLM performance, and implementing advanced training methodologies such as Supervised Fine\-Tuning (SFT), Direct Preference Optimization (DPO), and Proximal Policy Optimization (PPO).

Key Responsibilities

**1\. AI Architecture \& Solution Design**

* Design and architect scalable AI and Generative AI solutions based on business requirements.
* Develop enterprise\-grade LLM\-based applications and intelligent AI systems.
* Define model selection, fine\-tuning strategies, and deployment architectures.
* Drive AI solution development from proof of concept to production.

**2\. LLM Fine\-Tuning \& Optimization**

* Implement and optimize LLM fine\-tuning methodologies.
* Work on Supervised Fine\-Tuning (SFT) to customize foundation models for specific use cases.
* Apply Direct Preference Optimization (DPO) to improve model alignment and response quality.
* Leverage Proximal Policy Optimization (PPO) for reinforcement learning\-based model optimization.
* Implement parameter\-efficient fine\-tuning techniques, including LoRA and QLoRA.
* Evaluate model performance and optimize accuracy, efficiency, and inference costs.

**3\. Knowledge Graphs \& GraphRAG**

* Design and implement Knowledge Graph\-based AI solutions.
* Develop GraphRAG architectures to enhance retrieval, contextual understanding, and reasoning.
* Integrate structured and unstructured data sources into knowledge\-driven AI applications.
* Build intelligent retrieval pipelines combining graph\-based retrieval and LLM capabilities.
* Optimize knowledge representation and retrieval strategies.

**4\. AI Engineering \& Implementation**

* Develop end\-to\-end AI workflows for model training, fine\-tuning, evaluation, and deployment.
* Integrate fine\-tuned LLMs into enterprise applications and business workflows.
* Establish model evaluation and performance monitoring frameworks.
* Collaborate with engineering and product teams to deliver scalable AI solutions.

Required Technical Skills

* Large Language Models (LLMs)
* Model Fine\-Tuning
* Supervised Fine\-Tuning (SFT)
* Direct Preference Optimization (DPO)
* Proximal Policy Optimization (PPO)
* Low\-Rank Adaptation (LoRA)
* Quantized Low\-Rank Adaptation (QLoRA)
* Knowledge Graphs
* GraphRAG
* Generative AI
* LLM Architecture \& Optimization
* AI Model Evaluation \& Deployment

Desired Candidate Profile

* Strong experience in AI/ML engineering, Generative AI, and LLM\-based solution development.
* Hands\-on experience with fine\-tuning open\-source or foundation language models.
* Practical knowledge of SFT, DPO, PPO, LoRA, and QLoRA techniques.
* Experience designing Knowledge Graphs and implementing GraphRAG solutions.
* Strong understanding of LLM training, optimization, retrieval, and inference workflows.
* Ability to translate complex business requirements into scalable AI architectures.
* Strong problem\-solving, technical leadership, and stakeholder communication skills.

Key Deliverables

* Design and implementation of scalable LLM\-based AI architectures.
* Successful fine\-tuning and optimization of foundation models.
* Implementation of SFT, DPO, PPO, LoRA, and QLoRA approaches.
* Development of Knowledge Graph and GraphRAG\-based retrieval solutions.
* Production\-ready AI solutions with measurable improvements in model performance
Share