Amit Roy

Software Engineer

Bengaluru, Karnataka, India2 yrs 5 mos experience

AI EnabledAI ML Practitioner

Key Highlights

Expert in Generative AI and NLP solutions.
Proven track record in fraud detection and document intelligence.
Skilled in architecting scalable AI workflows.

Stackforce AI infers this person is a Healthcare-focused AI Engineer with expertise in Generative AI and NLP.

Contact

Skills

Core Skills

Generative AiAi WorkflowsNatural Language Processing (nlp)

Other Skills

LangGraphReActAPI automationcontext cachingclinical data analysisprompt engineeringLLMdocument intelligenceModel BuildingRAG techniquesKalman filtersdeep learningdata processingPEFT techniquesChromaDB

About

I am an AI Engineer with a passion for solving real-world problems through innovative AI and machine learning solutions. At Centific, I have worked on exciting projects like developing a Generative AI-powered resume fraud detection system, creating advanced computer vision models for object tracking and automating prompt and response validation process by fine tuning LLMs for large language models. With a solid foundation in AI from IIIT Allahabad and hands-on experience in tools like TensorFlow, PyTorch, and Azure, I thrive on building systems that are not only efficient but also impactful. Whether it is improving forecasting models, fine tuning LLMs or optimizing algorithms, I am always up for a challenge.

Experience

2 yrs 5 mos

Total Experience

1 yr 3 mos

Average Tenure

1 yr 1 mo

Current Experience

Concertai

Software Engineer - Gen AI

May 2025 – Present · 1 yr 1 mo · Bengaluru · On-site

1. Architected and deployed MCP-driven multi-agent systems using LangGraph and ReAct agents for clinical data analysis, enabling dynamic tool orchestration and scalable AI workflows
2. Built an agentic text-to-SQL pipeline to convert natural language queries into real-time medical dashboards, integrating prompt engineering, API automation, and context caching
3. Developed LLM-powered document intelligence systems to extract and structure information from medical prescriptions
4. Implemented context caching strategies to optimize LLM performance by reducing redundant computation and token usage
Key Impact:
1. Achieved 67% reduction in latency and 98% decrease in token usage in multi-agent workflows
2. Delivered 62% faster query processing and 90% cost savings in production text-to-SQL systems
3. Enabled scalable, production-ready AI solutions for clinical decision support and data analysis

LangGraphReActAPI automationcontext cachingGenerative AIAI workflows

Centific

AI Engineer

Jan 2024 – May 2025 · 1 yr 4 mos · Hyderabad · On-site

1. Developed an agentic RAG chatbot with integrated web search to extract insights from large-scale document sets (500+ pages)
2. Built a Generative AI-based fraud detection system to identify multi-identity submissions across 350,000+ resumes
3. Collaborated with cross-functional global teams to design LLM evaluation and validation pipelines, leveraging PEFT techniques (LoRA/QLoRA)
4. Engineered a multi-camera tracking system combining Kalman filters and deep learning (ResNeXt-50 with Triplet Loss) for identity recognition
Key Impact:
1. Improved retrieval accuracy by reducing ~30% irrelevant results using adaptive and corrective RAG techniques
2. Scaled fraud detection pipelines for high-volume, real-time processing on Azure
3. Achieved 88.3% IDF1 score on AI City 2023 multi-camera tracking benchmark
4. Enhanced LLM output quality through automated validation and fine-tuning workflows