AI Engineer – Generative & Agentic Workflows

• Model Optimization & Fine-Tuning: Fine-tune, distill, and optimize open-source and

proprietary foundation models (e.g., Llama, Claude, GPT-4, Mistral) using
techniques like PEFT/LoRA and RLHF/DPO.

• Retrieval-Augmented Generation (RAG): Build scalable, low-latency Hybrid RAG architectures

integrating vector databases, graph databases, and semantic routing for
complex domain contexts.

• Multimodal Pipelines:

Design and deploy multimodal pipelines (text, vision, audio, structured
data) to extract insights and generate rich artifacts.

Agentic
AI & Autonomous Systems

• Multi-Agent Architectures: Design and deploy multi-agent orchestration frameworks

(e.g., LangGraph, AutoGen, CrewAI, Semantic Kernel) with dedicated roles,
shared memory, and cross-agent negotiation strategies.

• Planning & Reasoning Protocols: Implement advanced reasoning paradigms—such as

Chain/Tree/Graph-of-Thought, ReAct loops, self-reflection, and
reflection-based error correction.

• Tool Augmentation & Function Calling: Connect agents to external APIs, databases, software

environments, and browser tools, enabling reliable function calling,
schema validation, and tool execution.

• Human-in-the-Loop (HITL) Workflows: Build human-in-the-loop safety checkpoints, approval

triggers, and oversight interfaces into autonomous agent loops.

Platform
Performance, Guardrails & MLOps

• Agentic Observability & Evaluation: Set up rigorous evaluation frameworks (LLM-as-a-judge,

trajectory tracing, latency profiling) using tools like LangSmith,
Phoenix, or Arize.

• Guardrails & Alignment: Implement strict safety, hallucination mitigation,

context-window optimization, and prompt injection defenses using guardrail
frameworks (e.g., NeMo Guardrails, Guardrails AI).

• Production Deployment: Scale agent workflows on cloud infrastructure

(AWS/Google Cloud Platform/Azure) with async task queues, durable execution state, and
low-latency API integration.

Key Skills & Technologies:

Languages
Python (Expert), TypeScript / Node.js (Plus)

Gen AI Frameworks
PyTorch, Hugging Face Transformers, vLLM, Ollama,
LangChain, LlamaIndex

Agentic Frameworks
LangGraph, AutoGen, CrewAI, Semantic Kernel, Temporal

Databases & Search
Qdrant, Pinecone, Milvus, Weaviate, Pgvector, Neo4j

Infrastructure & MLOps
Docker, Kubernetes, Ray, FastAPI, LangSmith, MLflow,
AWS/Google Cloud Platform/Azure

Experience
& Qualifications

• Experience: 6+ years of professional software engineering experience, with 2+ years

dedicated to building and deploying Gen AI and/or LLM applications in
production.

• Proven Track Record:

Experience building stateful LLM applications, custom RAG systems, or
autonomous agentic workflows deployed to real users.

• Strong Algorithmic Foundation: Solid understanding of transformer architectures,

attention mechanisms, vector embeddings, and non-deterministic state
machine design.

• Problem-Solving Mindset: Comfort dealing with model non-determinism, edge cases

in tool calling, and designing robust fallback mechanisms.