Looking for a job

Akshay Jyani

@akshayjyani

Senior Machine Learning Engineer building production LLM and on-device AI pipelines that deliver measurable performance.

India

Message

What I'm looking for

I want to build and deploy production ML systems—VLMs, LLMs, RAG pipelines, computer vision-based or NLP-based solutions—where performance is measurable, engineering is end-to-end, and I can optimize models for speed, size, and real user outcomes.

I, too, have experience in designing the data and also worked on the Synthetic Data Generation setup.

I’m a Senior Machine Learning Engineer focused on taking advanced ML from prototype to production—especially LLM and on-device AI systems. I built an end-to-end LLM pipeline for K-12 curriculum and assessment generation using agentic workflows with validation loops, post-trained custom models, and achieved 99%+ average quality on independent multi-dimensional evaluation (factual accuracy, curriculum alignment, difficulty calibration, and distractor quality).

At Samsung R&D Institute India (SRIB), I architected and deployed a production-ready Vision-Language Model pipeline for Galaxy devices using an x8 quantized, llama.cpp-powered inference engine (LLaMA C++), delivering real-time performance and up to 3× throughput speedup. I also designed high-accuracy, lightweight on-screen classification solutions (97.1%), built a compact CLIP model via knowledge distillation and pruning (40% size reduction, 90% accuracy), and shipped mobile screenshot barcode/QR detection with 92.7% accuracy—along with an AI WatchCrop solution reaching 92% user satisfaction.

Experience

Work history, roles, and key accomplishments

Machine Learning Engineer

Incept Labs

Jan 2026 - Jun 2026 (5 months)

Built an end-to-end LLM pipeline for K-12 curriculum and assessment generation using agentic workflows with validation loops, plus custom model post-training and fine-tuning. Deployed production ML services on AWS using SageMaker, ECS/ECR, Lambda, SQS, API Gateway, and CloudWatch.

Prompt Engineering Model Fine Tuning QLoRA AWS Sagemaker Amazon ECS Amazon ECR AWS Lambda Amazon SQS API Gateway Amazon S3 Cloudwatch Arabic LLMs SFT Multi Step Generation

Senior Machine Learning Engineer

Samsung R&D Institute India

Aug 2023 - Jan 2026 (2 years 5 months)

Architected and deployed a production-ready vision-language model (VLM) pipeline with quantized inference for Samsung Galaxy S25 devices, achieving real-time performance and throughput improvements. Built on-device classification and lightweight CLIP-based models, and delivered applications including Sticker Search and WatchCrop.

Vision Language Models Llama.Cpp LLaMA C++Quantization (X8)Grammar Based Shared Objects CNN Classifiers Transformers Knowledge Distillation Pruning Lightweight Model Design Stable Diffusion Web Scraping Synthetic Data Generation Saliency Models Mobile ML Deployment LLamacpp

Software Developer Intern

Samsung R&D Institute India

May 2022 - Jul 2022 (2 months)

Contributed to a phonetic search feature to retrieve similar-sounding Indian names across 7+ languages, including dataset creation and on-device deployment. Worked with neural phonetics concepts such as seq2seq architectures using LSTM and time-distributed layers to support model-aligned dataset design.

Multilingual (Indian Names)Neural Phonetics Time Distributed Layers On Device Deployment