Agentic AI & Orchestration
Built LangGraph workflows with specialized retrieval, reasoning, validation, and escalation agents, backed by PostgreSQL/pgvector.
Senior Generative AI Engineer / Agentic AI & LLM Systems
Ahamed Musthafa R S
I build agentic AI and production LLM systems for enterprise customer support and life sciences. My work spans multi-agent orchestration, RAG, fine-tuning, and reliable AI workflows, backed by Python, FastAPI, Django, PostgreSQL, and AWS.
- 7+
- Years in software
- 3+
- Years in LLMs
- 75%
- Faster query-builds
- ~50%
- Less integration time
About
Enterprise AI with grounded outputs and controlled workflows.
I am Ahamed Musthafa R S, a Senior Generative AI Engineer with 7+ years of software engineering experience and 3+ years building production GenAI and LLM systems. Since August 2026, I have been working at Tredence Inc in Bengaluru on retrieval-augmented multi-agent support workflows.
My experience spans enterprise customer support and life sciences: specialized agents, grounded-source validation, compliance checks, clinical copilots, and domain-specific fine-tuning. At Saama, my work reduced clinical query-build time by 75%, model integration time by about 50%, and platform effort by up to 80%.
Fine-Tuned Models
Fine-tuned open-source LLMs with LoRA on synthetic domain datasets, achieving over 80% classification accuracy for logical discrepancies in clinical data.
Reliable LLM Workflows
Built an orchestration harness for controlled execution of deterministic AI workflows, with grounded-source validation, compliance checks, and workflow controls.
Experience
-
Aug 2026 - Present
Bengaluru, KASenior Gen AI Engineer Tredence Inc
Developed a retrieval-augmented multi-agent support workflow using LangGraph and PostgreSQL/pgvector, grounding responses in enterprise knowledge with dedicated retrieval, reasoning, validation, and escalation agents.
-
Jun 2021 - Jul 2026
Coimbatore, TNSenior Research Engineer Saama Technologies
Built controlled LLM orchestration, an Interactive Review Listing copilot, and natural-language code generation tools. Fine-tuned domain LLMs with LoRA and designed model service interfaces for production rule engines.
-
Jul 2019 - Jun 2021
Bengaluru, KABackend Developer Zelthy
Built pharma CRM and therapy-management applications with Python, Django, JavaScript, and PostgreSQL. Implemented patient-support program rules that improved product efficiency by 40%, and added AWS CloudWatch monitoring.
Education
B.E. in Computer Science and Engineering, Government College of Technology, Coimbatore 2015 - 2019
GenAI Stack
Agents, retrieval, and production LLM engineering.
Specialized agents, grounded enterprise responses
At Tredence, I developed a retrieval-augmented multi-agent support workflow using LangGraph and PostgreSQL/pgvector, with specialized agents for retrieval, reasoning, response validation, and escalation.
Approach
Multi-agent orchestration
Use
Enterprise support
Evidence
Tredence workflow
boot: genai portfolio online
- GenAI
- OpenAI, Cohere, LangChain, LangGraph, Agentic AI, RAG, Prompt Engineering, LLM Evaluation, LoRA, QLoRA, Transformers
- Backend
- Python, FastAPI, Django, Flask, SQLAlchemy, Celery, JavaScript, SQL, NoSQL
- Cloud & Data
- AWS EC2, S3, RDS, Lambda, Bedrock, CloudWatch, Boto3; PostgreSQL/pgvector, MySQL, MongoDB, Snowflake; Docker, kubectl, Airflow
- Tools & Domain
- Gradio, Streamlit, Pandas, Keras, Scikit-learn, Selenium, BeautifulSoup, Plotly, Postman, Windsurf, Claude; CDISC standards, clinical trial design, pharma CRM
Selected Work
GenAI builds from product systems to personal experiments.
Clinical AI Workbench
A lightweight, no-PHI/no-PII pharmacovigilance workbench for public adverse-event exploration. Combines PRR/ROR signal metrics and source-grounded summaries with citation-first retrieval from openFDA, FDA labels, RxNorm, and PubMed.
Live WorkbenchDeciSentient
An autonomous decision agent for shopping and travel. TinyFish WebAgent handles web interaction, Gemini supports reasoning, and ElevenLabs enables conversational voice research and comparison.
RepositoryCodeProof
An autonomous code-review and remediation agent that analyzes repositories, detects fragile implementations and technical debt, reasons about corrective actions, and generates targeted code changes.
Contact