AI / ML Engineer

We are looking for a passionate and technically strong AI/ML Engineer to join our team.

This is a hands-on engineering role with a research orientation. You will work on fine-tuning open-source models, designing machine learning pipelines, and deploying production-ready AI systems at scale.

Our key focus areas include:

• Handwritten answer evaluation

• Marathi Automatic Speech Recognition (ASR)

• Computer Vision for Indic scripts

• Low-resource language AI models

• Efficient cloud and on-device AI inference

What makes this role exciting?

We have already shipped a fine-tuned vision model (Gemma 4 31B using Unsloth and LoRA) that surpasses many state-of-the-art models on Marathi children's handwriting recognition.

You will build on this foundation, improve existing systems, and explore new approaches to solving challenging AI problems.

Important: This is not a data annotation or prompt engineering role. We are looking for someone who understands how machine learning models work internally.

Key Responsibilities

1. Model Fine-Tuning & Training

• Fine-tune open-source LLMs and vision models using LoRA, QLoRA, and other Parameter-Efficient Fine-Tuning (PEFT) techniques.

• Work with Hugging Face Transformers, Unsloth, and PEFT libraries.

• Optimize GPU training performance and resource utilization.

• Experiment with model architectures for domain-specific educational applications.

2. Computer Vision & Speech AI

• Develop models for Marathi handwriting recognition and answer evaluation.

• Work on speech recognition systems for Marathi and other Indic languages.

• Design AI-based rubric scoring and evaluation mechanisms.

• Handle noisy, inconsistent, and low-resource real-world datasets.

3. Model Deployment & Optimization

• Deploy fine-tuned models to AWS infrastructure, including ECS/Fargate and GPU inference endpoints.

• Optimize inference performance, latency, throughput, and infrastructure cost.

• Explore quantization, batching, ONNX, TensorRT, and other inference optimization techniques.

• Investigate on-device AI deployment for Android applications.

4. ML Pipelines & Evaluation

• Build end-to-end ML pipelines covering data preprocessing, augmentation, training, evaluation, and serving.

• Define and track evaluation metrics such as Word Error Rate (WER), Character Error Rate (CER), and Devanagari-aware benchmarks.

• Perform structured error analysis to identify model weaknesses.

• Improve model reliability and accuracy through systematic experimentation.

5. Research & Engineering Collaboration

• Read and evaluate recent AI/ML research papers.

• Translate research ideas into practical experiments and production implementations.

• Collaborate with backend engineers to integrate AI models into FastAPI services.

  • Contribute to technical decisions regarding model architecture, training, and deployment.