Akash Panzade

AI Researcher

Hyderabad, Telangana, India4 yrs 10 mos experience

AI ML PractitionerHighly Stable

Key Highlights

Expert in building scalable AI solutions.
Proven track record in healthcare automation.
Strong background in MLOps and deployment.

Stackforce AI infers this person is a Healthcare-focused Data Scientist with strong expertise in AI and MLOps.

Contact

akashpanzade2000@gmail.com LinkedIn

Skills

Core Skills

Machine LearningGenerative AiNatural Language Processing

Other Skills

AI AgentsAI securityAlgorithmsAmazon EKSAmazon Web Services (AWS)Architectural AnalysisArtificial Intelligence (AI)BERTBig Data AnalyticsBusiness RequirementsC (Programming Language)C++CI/CDCode OptimizationCompiler Design

About

Data Scientist & ML Engineer | IIT Kanpur CSE 🚀 I specialize in researching, developing, and deploying high-performance ML, DL, & LLM-based systems, transforming unstructured data into scalable, production-grade AI solutions that deliver measurable business impact. 💡 Built intelligent systems that power healthcare automation, large-scale forecasting, NL2SQL translation, efficient ML & RecSys systems, and LLM-based assistants leveraging models like LLaMA, GPT, and BERT, fine-tuned using QLoRA, PEFT, and orchestrated via RAG, vector databases, multi-agent systems, and knowledge graph generation, etc. My work bridges applied research and engineering to deliver explainable, contextual, and high-performing GenAI pipelines. 🌐 I bring hands-on expertise in building end-to-end solutions from model development and evaluation to scalable inference, real-time APIs, backend integration, and cloud deployment across AWS, GCP, and hybrid infrastructures. I've implemented scalable CI/CD workflows (GitHub Actions, Terraform), MLOps/LLMOps best practices, and experiment tracking (MLflow, Weights & Biases, LangSmith, PromptLayer, OpenAI Evals, etc) to ensure robustness, reproducibility, and low-latency performance in production. 🌱 Continuously evolving with advancements in AI/ML/GenAI, I’m driven to build innovative, dependable, and impactful AI systems. 🔍 Currently open to full-time roles in AI/ML, Data Science, or Generative AI, where I can contribute to building reliable, scalable, and production-ready intelligent systems. 📧 Let’s Connect: akashpanzade2000@gmail.com.

Experience

4 yrs 10 mos

Total Experience

2 yrs

Average Tenure

10 mos

Current Experience

Teradata

AI/ML Engineer

Aug 2025 – Present · 10 mos · Hyderabad, Telangana, India

Carelon global solutions

Data Scientist

Jul 2023 – Aug 2025 · 2 yrs 1 mo · Hyderabad, Telangana, India · Hybrid

🤖 AI/ML
Built XGBoost classifier with BERT & Word2Vec embeddings to automate ticket triage, saving 8–10 FTEs & standardizing workflows.
Engineered probabilistic high-cost claimant predictor using explainable AI for proactive risk management & cost optimization.
Optimized SARIMA & Prophet models, achieving record-low RMSE of 9.05 for claims forecasting.
Developed clinical recommendation systems driving underwriting & treatment improvements.
Improved underwriting precision with explainable ML models, enhancing clinical risk scoring accuracy.
🧠 GenAI & LLM
Built medical QA system using LLaMA-2, RAG, LangChain, and QLoRA 4-bit quantization, reducing inference costs &improving accuracy.
Developed scalable NL-to-SQL system with fine-tuned LLMs, semantic search, and automated KGs, achieving 95% accuracy, <5s latency.
Enabled proactive CoC platform with GenAI forecasting on millions of data points for predictive state-level plans.
Expanded member outreach from 48% → 81% with LLM-powered targeted interventions.
Designed modular GenAI systems (agentic memory, recursive planning) using LangChain & LangGraph.
Fine-tuned domain-specific LLMs surpassing GPT on clinical entity recognition and specialized tasks.
🚀 MLOps & Deployment
Led End-to-End Deployment of ML/LLM solutions across hybrid infrastructures (AWS, GCP, on-premises), leveraging Docker, CI/CD, and MLOps to build robust, production-grade systems.
Built high-performance inference APIs & backend systems converting AI insights into business value.
Ensured model reliability with automated monitoring, retraining, & data drift management.
Developed A/B testing frameworks for data-driven model optimization.
🏆 Recognition & Awards
Advanced GenAI-based Analytics Solutions Award for AI-driven clinical review transformation.
Rapid Product Deployment Award for accelerating mission-critical AI product launch.
3x Go Above Award for exceptional dedication & innovation in data science.

XGBoostBERTWord2Vecexplainable AISARIMAProphet+10

Indian institute of technology, kanpur

2 roles

Teaching Assistant

Jul 2021 – Apr 2023 · 1 yr 9 mos

ESC101 – Fundamentals of Computing (Jul 2021 – Jul 2022)
CS610A – Programming for Performance (Aug 2022 – Dec 2022)
CS335A – Compiler Design (Jan 2023 – May 2023)
Mentored students across core CS courses, guiding them through advanced algorithms, systems architecture, and compiler design; fostered strong conceptual understanding and practical problem-solving through hands-on, application-focused instruction.

Graduate Research Assistant

Jun 2021 – May 2023 · 1 yr 11 mos

🔬 Applied AI/ML Research | Security | NLP & GenAI
Conducted end-to-end research in AI security, NLP, and knowledge representation, driving both publication and prototype development.
Developed an ML-powered ransomware-detection framework that combines behavioral signals with cryptographic heuristics; paper published at IEEE.
Designed transformer-based NLP pipelines for real-time text classification and tagging, incorporating model compression for on-device inference and achieving robust multilingual performance.
Built scalable knowledge-graph generation and entity-linking modules that feed downstream semantic-search and information-retrieval systems, improving query relevance and recall.
Collaborated on graph-based threat-intelligence modeling and multi-label classifiers, enabling real-time detection and triage of cybersecurity incidents.

AI securityNLPknowledge representationransomware detectiontransformer-based NLP pipelinesknowledge-graph generation+4