Senior Software Engineer - Neuro-Symbolic AI & LLMs

About The Role

As a Senior Software Engineer for a platform that creates small/compact, custom/target and neurosymbolic AI models for Enterprises, you will lead the architectural development and implementation of hybrid AI systems. You will take a hands on role in building and leveraging knowledge bases, going beyond off the shelf LLMs to train, fine tune, and optimize large scale models that work in tandem with structured knowledge systems.

About The Role

As a Senior Software Engineer for a platform that creates small/compact, custom/target and neurosymbolic AI models for Enterprises, you will lead the architectural development and implementation of hybrid AI systems. You will take a hands on role in building and leveraging knowledge bases, going beyond off the shelf LLMs to train, fine tune, and optimize large scale models that work in tandem with structured knowledge systems.

About The Role

As a Senior Software Engineer for a platform that creates small/compact, custom/target and neurosymbolic AI models for Enterprises, you will lead the architectural development and implementation of hybrid AI systems. You will take a hands on role in building and leveraging knowledge bases, going beyond off the shelf LLMs to train, fine tune, and optimize large scale models that work in tandem with structured knowledge systems.

Key Responsibilities

• Neuro Symbolic Integration: Design and implement architectures that link LLMs with Knowledge Graphs (KGs) to ensure factual grounding and commonsense reasoning.


• LLM Training & Optimization: Lead the training and fine tuning of Large Language Models (LLMs) on domain specific datasets (e.g., healthcare, pharmaceuticals) using techniques such as LoRA, QLoRA, and distributed training.


• Knowledge Graph Engineering: Build and maintain scalable knowledge representation systems, ontologies, and triple stores that serve as the truth layer for AI agents.


• System Architecture: Develop high performance backends for Neuro Symbolic AI systems, ensuring low latency inference and seamless integration between neural and symbolic components.


• Research to Production: Translate cutting edge research papers on C3AN methodologies into robust, production ready code.


• Mentorship: Lead a team of engineers and researchers, fostering a culture of rigorous engineering and innovative AI development.

Required Skills & Experience

• Education: M.S. or PhD in Computer Science, AI, or a related field (or equivalent industry experience).


• Core Engineering: 5+ years of software engineering experience with deep proficiency in Python.


• LLM Expertise: Extensive experience in Large Language Model training, including pretraining, instruction tuning, and RAG (Retrieval Augmented Generation) architectures.


• Symbolic AI: Strong foundation in Knowledge Graphs, SPARQL/Cypher, RDF/OWL, and logic based reasoning frameworks.


• ML Frameworks: Expertise in PyTorch or TensorFlow, along with experience using distributed training libraries such as DeepSpeed or Megatron LM.


• Cloud & Infrastructure: Familiarity with GPU orchestration (CUDA, Triton) and deploying models on high performance compute environments (e.g., IndiaAI Cloud).

Bonus Qualifications

• Experience with Neuro Symbolic frameworks, Knowledge Graphs, and logical reasoning systems (e.g., Logic Tensor Networks, DeepProbLog).


• Contributions to open source AI projects or publications in leading venues such as NeurIPS, ICML, or AAAI.


• Experience in domains such as manufacturing, pharmaceuticals, or healthcare.

About Project C3AN

C3AN is a flagship project focused on the next generation of AI: Custom (solving specific tasks or domain specific problems), compact (small or right sized), Neuro Symbolic AI. The project aims to address the limitations of purely data driven models such as hallucinations and limited reasoning by integrating Large Language Models (LLMs) with Symbolic AI (Knowledge Graphs, ontologies, and logic systems).

The goal is to build context aware, explainable, and factually grounded systems that leverage both "System 1" (fast, intuitive) and "System 2" (deliberate, logical) modes of reasoning.