Abhishek Dasgupta — AI Researcher

i'm a 23-year-old cs engineering graduate who’s been building from day one. i began my professional journey in my first year - not in tech, but as a content writing intern, just to get my foot in the door and learn how to earn money from skills. from second year onwards, i found my space in generative ai at a time when chatgpt just launched and have since interned at 4 early-stage startups as an AI Engineer across sales, edtech, healthcare, and devtools - applying generative ai and llms to real-world problems. right now, I’m working as an AI Engineer at Zykrr, where i design and build AI-powered analytics systems on top of large-scale customer survey data. i developed an enterprise cx chatbot that sits directly on top of structured and unstructured feedback datasets, enabling teams to query thousands of survey responses in natural language and extract actionable insights using LLM-powered retrieval and reasoning pipelines. i also architected the AI service layer for our case management product from scratch building pipelines that automatically convert raw survey feedback into structured issues, classify intent and sentiment, and route them to the appropriate departments. previously, i worked as a software engineering intern at raccoon ai, where i was in a full-stack generalist role - writing golang backend apis, building frontend components with nextjs, automating browsers with playwright, and integrating llms to create an autonomous browser agent that can perform web tasks through natural language. working across multiple startups has taught me how to take ownership, solve problems independently, and adapt quickly. i enjoy working in fast-paced teams where learning is constant and every line of code has real impact. i'm passionate about solving meaningful problems with applied ai, building things from scratch, and working with people who move fast, learn faster, and build with purpose. you can reach me at abhishek20dgp@gmail.com

Stackforce AI infers this person is a SaaS-focused AI Engineer with expertise in Generative AI and cloud technologies.

Location: Gurugram, Haryana, India

Experience: 3 yrs 7 mos

Skills

Artificial Intelligence (ai)
Devops
Software Development

Career Highlights

Led development of AI features at multiple startups
Expert in Generative AI and Large Language Models
Strong foundation in DevOps and cloud infrastructure

Work Experience

Zykrr — The CX Monetization Company

AI Engineer (10 mos)

Raccoon AI

Software Engineer Intern (Fullstack) (6 mos)

Tap Health

Generative AI Intern (4 mos)

Ground0

Generative AI Intern (6 mos)

GDSC BCREC

GDSC Lead (1 yr 1 mo)

AI/ML Lead (9 mos)

TheSales.io

Machine Learning Intern (10 mos)

AccioIbis

Machine Learning Intern (1 yr 1 mo)

iCoders

Core Member (1 yr)

Python Mentor (7 mos)

Mobile Mode Gaming

Content Writer (1 yr)

Education

Bachelor of Technology - BTech at Dr. B.C. Roy Engineering College

CBSE at D.A.V. Model School, Durgapur

10th at St. Xavier's School Durgapur

Abhishek Dasgupta

AI Researcher

Gurugram, Haryana, India3 yrs 7 mos experience

AI EnabledAI ML Practitioner

Key Highlights

Led development of AI features at multiple startups
Expert in Generative AI and Large Language Models
Strong foundation in DevOps and cloud infrastructure

Stackforce AI infers this person is a SaaS-focused AI Engineer with expertise in Generative AI and cloud technologies.

Contact

Skills

Core Skills

Artificial Intelligence (ai)DevopsSoftware Development

Other Skills

AI EngineerNatural Language Processing (NLP)Large Language Models (LLM)KubernetesAWSSQLAIPython (Programming Language)Go (Programming Language)Next.jsplaywrightTypeScriptMongoDBApache KafkaGenerative AI

About

Experience

3 yrs 7 mos

Total Experience

1 yr 8 mos

Average Tenure

10 mos

Current Experience

Zykrr — the cx monetization company

AI Engineer

Jul 2025 – Present · 10 mos · Gurugram · Hybrid

At zykrr we are building ai-powered surveys and analytics that help you understand your customers faster.
I lead the development of the platform’s ai features while also having the additional role of designing and maintaining the devops infra that deploys and runs all zykrr services.
Built the foundation of Zyva, a multi-agent conversational CX analytics platform that orchestrates specialized LLM agents for natural-language query interpretation, NL to SQL generation, analytics computation, prompt-to-chart visualization, and theme clustering, enabling CX leaders to explore large-scale survey datasets through conversational queries.
Developed a Natural Language to SQL engine that translates user queries into optimized analytical SQL over a ClickHouse data warehouse, enabling dynamic CX analytics including NPS trends, response analysis, and insight generation.
Architected the AI service layer for a case management system that transforms unstructured survey feedback into structured issues with categories, priorities, metadata enabling automated routing of cases to the appropriate teams.
Devops:
Designed and deployed the K8 infra for the Phoenix platform on AWS EKS, configuring NGINX Ingress Controller, K8 services, API Gateway integration, and pod/service manifests to manage and expose microservices across environments.
Built the entire CI/CD pipeline from scratch for phoenix project including frontend, backend, AI and auth services using GitLab CI, Docker, and Kubernetes, where tagged releases automatically trigger container builds, push images to the GitLab Container Registry, and deploy updated services to EKS through manifest-based kubectl deployment scripts.
Implemented automated deployment and secrets management, where deployment scripts fetch encrypted configuration from AWS S3 and populate K8 Secrets, overall reducing the manual build and deployment workflow from ~1 hour to a two-click release process (test and production environments).

AI EngineerNatural Language Processing (NLP)Large Language Models (LLM)KubernetesAWSSQL+2

Raccoon ai

Software Engineer Intern (Fullstack)

Jan 2025 – Jul 2025 · 6 mos · Bengaluru, Karnataka, India · On-site

worked on building an ai-powered autonomous browser agent capable of navigating the internet, performing actions across websites, and extracting data, available both as an api and a chat-based product.
key contributions:
designed and improved llm-based logic to convert natural language commands into browser actions, making the agent more intelligent and flexible.
implemented real-time polling in golang to fetch task updates from mongodb, and integrated apache kafka for streaming live task results through a producer-consumer system.
built and integrated tools using python and playwright for automating browser actions and domain-specific search flows.
contributed major enhancements across the system, improving dom interaction (dropdowns, radio buttons, text inputs), session control, media playback, and chat consistency.
enabled observability with langsmith tracing and developed automated test scripts for core tools to improve reliability and debugging.
created a video livestreaming interface using websockets to visualize browser activity using next.js and typescript, and redesigned the platform page to better showcase live agent sessions.
improved agent robustness by implementing stealth browsing (dns/webrtc leak prevention), session pause/resume, and captcha detection and handling.
collaborated with the team to enhance llm-driven task orchestration, middleware logic, and api structure for more reliable autonomous execution.

Python (Programming Language)Go (Programming Language)Next.jsLarge Language Models (LLM)playwrightTypeScript+3

Tap health

Generative AI Intern

Sep 2024 – Jan 2025 · 4 mos · Gurugram, Haryana, India · Remote

Worked on R&D of the proprietary VAID-2 AI Diagnosis Model, designed to analyze user symptoms and provide accurate diagnoses through adaptive follow-up questioning.
Developed several AI inference APIs using FastAPI for the backend team.
Integrated medical knowledge bases leveraging ChromaDB, Neo4j Graph Database, and MongoDB, utilizing a HybridRAG architecture for efficient and precise data retrieval.
Developed the Instant Answers V2 module APIs for the Tap Health App.
Contributed to the integration of a consumer-facing AI chatbot for a hospital, enhancing patient engagement
Participated in regular meets with the Product Team and Medical Team for planning and building different modules of the app.
Managed and maintained deployments on Azure VMs.
Conducted comparative testing and evaluation of open-source LLMs in the medical domain.

Artificial Intelligence (AI)Generative AILarge Language Models (LLM)FastAPIMicrosoft Azure

Ground0

Generative AI Intern

Feb 2024 – Aug 2024 · 6 mos · Mumbai · Remote

Successfully built an MVP for the company through Generative AI integration for an AI Career Co-Pilot using Google Gemini, Milvus Vector Database, Flask, Strapi.
Developed efficient Retrieval -Augmented-Generation pipelines using Langchain, embedding models and vector databases to fetch similar career roles based on user profiles.
Used LangGraph for evaluating prompts and RAG Pipelines.
Conducted in-depth research and delivered consultancy on GenAI solutions for the company’s core products.

Artificial Intelligence (AI)Large Language Models (LLM)Retrieval-Augmented Generation (RAG)Python (Programming Language)FinetuningGoogle Gemini

Gdsc bcrec

2 roles

GDSC Lead

Jul 2023 – Aug 2024 · 1 yr 1 mo

Got selected by Google Developer Student Clubs India team to lead the GDSC Club at our campus.

AI/ML Lead

Oct 2022 – Jul 2023 · 9 mos

Thesales.io

Machine Learning Intern

Feb 2023 – Dec 2023 · 10 mos · Île-de-France, France · Remote

Worked here through Acciolbis

Accioibis

Machine Learning Intern

Nov 2022 – Dec 2023 · 1 yr 1 mo · Paris, Île-de-France, France

Developed an AI-powered chatbot from scratch, utilizing Flask, OpenAI, Langchain, MongoDB, Kubernetes, and Google Cloud Platform, achieving over 80% faster performance through RAG, document retrieval, and streaming architecture enhancements.
Led research and development efforts in fine-tuning LLMs, notably Vicuna and Llama, outperforming ChatGPT benchmarks.
Created an AI bot for automating sales calls, significantly streamlining the sales process through intelligent, automated customer interactions.
Worked with TTS models, including faster whisper and distil whisper, significantly reducing latency for improved user interaction.
Designed a multifunctional AI assistant capable of customized creation, executing functions, performing custom actions, and facilitating internet scraping and image generation, broadening the scope of AI applications.

Stable DiffusionPython (Programming Language)TransformersPyTorchGenerative AILarge Language Models (LLM)+9