Open to opportunities

Amit Singh

@amitsingh18

Message

AI Software Engineer building production agentic multimodal systems across healthcare, enterprise, and media.

India

Message

What I'm looking for

I’m looking for a team where I can build production agentic, multimodal AI systems—shipping scalable pipelines, low-latency user experiences, and clear quality gains (latency, accuracy, reliability) from day one.

I build production-grade agentic AI systems and multimodal applications that feel fast, reliable, and practical. At InspironLabs, I architected a LangGraph-based multi-agent orchestration layer with parallel task execution (dependency wave calculation) to run 10+ concurrent tasks across SQL, RAG, and API datasources.

I also focus on measurable quality improvements end-to-end: I engineered a production RAG pipeline (Redis caching, Elasticsearch vector search, citation tracking, semantic chunking for 100+ page PDFs) and added a DSPy-driven conversation validation layer that reduced failed queries by 40%. In healthcare, I deployed a real-time voice + chat agent using LiveKit (WebRTC SFU) for sub-second responses, and I’ve complemented agent work with scalable data/ETL and ML foundations from Airflow/Snowflake pipelines and RLHF fine-tuning.

Experience

Work history, roles, and key accomplishments

Current

AI Software Engineer

Current

InspironLabs

Feb 2025 - Present (1 year 3 months)

Architected a LangGraph-based multi-agent orchestration system enabling concurrent execution of 10+ tasks across SQL, RAG, and API data sources, and built a production RAG pipeline processing 100+ page PDFs with citation tracking. Reduced failed queries by 40% via a conversation validation layer and delivered real-time multimodal voice/chat healthcare workflows with sub-second response times.

LangGraph DSPy LiteLLM Kafka Neo4j Elasticsearch RAG WebSockets GraphQL LiveKit

Machine Learning Engineer

PyVision Technologies

Jan 2024 - Jan 2025 (1 year)

Fine-tuned large language models using RLHF to improve alignment and response quality through reward modeling and human feedback loops. Contributed to optimization of reward models and training pipelines to support scalable model evaluation and improvement.

Reward Optimization Model Fine Tuning LLM Evaluation Optimization Embeddings

Data Analyst Intern

Hitachi Vantara

Jan 2023 - Aug 2023 (7 months)

Performed EDA on large-scale sensor data using PySpark and built ETL pipelines to automate ingestion and processing, enabling identification of patterns and anomalies. Enhanced Grafana dashboards by integrating additional sensor data to improve monitoring and reporting visibility.

Pyspark EDA ETL Data Ingestion Grafana Dashboards Sensor Data