Embedded AI Engineer

Job Title: Embedded AI Engineer

About the Role

We are looking for an innovative and technically strong Embedded AI Engineer to join our R&D division at Talrop. The ideal candidate will design, optimize, and deploy advanced AI models—particularly large language models (LLMs)—on embedded and edge devices. You'll work at the intersection of machine learning, embedded systems, and robotics, developing intelligent control and perception capabilities for next-generation autonomous drones and robotics platforms.

Key Responsibilities

Benchmark, select, and optimize LLMs (e.g., Google Gemini Nano, Llama, Qwen) for real-time deployment on edge devices such as Jetson Orin/Nano.

Quantize and compress models (INT8/FP16) while tuning inference frameworks (TensorRT, vLLM, Ollama) to minimize latency and power usage for drone applications.

Design, implement, and fine-tune NLP pipelines that transform natural language commands into actionable drone instructions (intent classification, safety validation).

Develop prototype systems enabling voice/text-based drone control, integrating real-time decision-making and flight automation.

Integrate AI models with embedded systems, ensuring safe and continuous communication between inference outputs and flight controller APIs.

Build monitoring pipelines for model performance, latency tracking, safety logging, and robustness validation.

Design and implement safety guardrails and simulation frameworks to validate all NLP-based commands before execution.

Manage model training workflows, prompt engineering, and inference benchmarking for field-specific applications.

Collaborate with firmware and robotics teams during field testing, analyze performance, and contribute to iterative AI improvements.

Document AI architecture, validation methods, and participate in regulatory and compliance documentation for autonomous systems.

Essential Skills

Programming Languages (Must Have)

Expert in Python (data structures, algorithms, and efficient coding).

Experience with CUDA or GPU programming.

Familiarity with C++ for performance-critical inference tasks.

Machine Learning Frameworks (Critical)

Proficient in PyTorch or TensorFlow.

Experience with Hugging Face Transformers and LangChain for LLM workflows.

Knowledge of ONNX for model conversion and optimization.

LLM-Specific Knowledge (Critical for This Role)

Deep understanding of Transformer architectures (GPT, BERT, T5, Llama, Qwen).

Experience in fine-tuning techniques (Full-tuning, LoRA, QLoRA).

Skills in prompt engineering, in-context learning, and few-shot learning.

Model Optimization & Compression (Essential for Edge)

Experience in quantization (INT8, FP16, INT4), pruning, and knowledge distillation.

Understanding of accuracy–latency–power trade-offs in edge AI.

Edge Deployment (Critical for Your Startup)

Experience optimizing models for Jetson Orin/Nano or similar devices.

Proficiency with TensorRT, vLLM, Ollama, , or other edge runtime frameworks.

Skills in latency benchmarking, power profiling, and real-time inference management.

MLOps & Production Systems

Model serving using FastAPI, Flask, or TorchServe.

Knowledge of Docker, Kubernetes, and CI/CD tools (MLflow, Weights & Biases, DVC).

Experience in model versioning, rollbacks, and continuous integration for ML pipelines.

Development Tools & Workflow

Strong with Git and version control for model/code management.

Proficient in Jupyter notebooks and Linux command-line tools.

Exposure to cloud ML platforms like AWS SageMaker, Azure ML, or Google Colab.

Good to Have

NLP and domain-specific processing: NER, intent classification, and voice/speech processing (Whisper, ASR).

Computer vision integration: multimodal models and real-time object detection.

Background in robotics, aerospace, or autonomous system design.

Qualification

Bachelor's degree in Computer Science, Mathematics, or Physics (minimum).

Master's degree in Computer Science, Machine Learning, or AI (preferred for senior roles).

PhD is valued but not mandatory with 4+ years of practical experience.

Equivalent profiles (e.g., coding bootcamp + 3+ years of production ML experience) are also considered.

Job Type: Full-time

Pay: ₹30, ₹50,000.00 per month

Ability to commute/relocate:

• Trivandrum, Kerala: Reliably commute or planning to relocate before starting work (Required)

Application Question(s):

• Are you willing to relocate to anywhere in Kerala?

Experience:

• relevant: 3 years (Required)

Work Location: In person