Four checks, every engagementRun in order
{{obC.nn}}{{obC.h}}{{obC.sr}}
{{obC.d}}
scroll, swipe or use the arrowsyour cursor is the laser
{{live}}
Punitkumar Harsur, AI & Data Enablement Consultant, Bengaluru, India
Complex ideas,
clearly made.
I'm Punit, an AI and data enablement consultant in Bengaluru. I help L&D and business teams turn data and GenAI goals into learning people can use.
A talk in eleven slides·scroll, swipe or use the arrows
01 / 11
Proof · where the years went
Seven years between data and learning.
Two bundles of strings, Data in teal and Learning in orange, twist around each other from 2019 to now. They cross at each role: research at the IoT Lab, UVCE, in 2019; Associate Data Science Consultant at UNext Learning in 2021; Product Strategy and Data Science Consultant at TimesPro in 2022; Consultant, AI and Machine Learning at Dono Consulting in 2026. Then they merge into one warm cable that runs into four numbers: 7 years between data and learning, 50+ trainings and executive education programmes, 3 years of Azure data engineering, 14 corporate and 7 academic organisations.
{{yrLabel}}
Delivered for GlobalFoundries, Honeywell, Deloitte, IBM, IIT Delhi and IIT Roorkee, via TimesPro and Dono Consulting.
{{bubble}}
02 / 11
How I work
Four steps, one slingshot.
A method, not a menu. Every engagement falls toward the same goal, gains speed at four stations, and leaves with the team running it themselves.
A probe falls toward a black hole along a dotted orange path, lights four stations on the way in, swings around the hole at its fastest point and is flung out to the right. Discover, Design, Build and deliver, Hand over, then the team runs it.
seven years in, still the same four moves
{{ob.capH}} {{ob.capK}}
03 / 11
Agenda · Selected work
Six stops: five cases, one in the making.
{{cur.n}} · {{cur.meta}}
{{cur.title}}
- {{p.h}}
- {{p.t}}
04 / 11
Case 01 · Dono Consulting · 2026
Enterprise AI & Automation Learning Labs
- Problem
- Business teams needed to use GenAI in real document and data work without skipping human review.
- What I did
- Scoped it with client L&D and business teams; designed workflows people build themselves with approval checkpoints, testing and responsible-AI guidance; ran walkthroughs and knowledge transfer.
- Outcome
- Client teams kept running the content in their own LMS.
- Reached people
- Through walkthroughs and knowledge transfer with the client teams.
05 / 11
Case 02 · TimesPro · 2022 to 2026
Executive education learning paths with IIT partners
- Problem
- Working professionals needed data and GenAI learning that matched industry practice.
- What I did
- Designed learning paths and case studies across data science, Azure data engineering, analytics and GenAI; used market benchmarking to keep the curriculum current.
- Outcome
- Delivered with IIT Delhi, IIT Roorkee and IIT Guwahati.
- Reached people
- Taught Azure, SQL, Python, PySpark, Power BI and EDA to working professionals.
06 / 11
Case 04 · TimesPro
Teaching grounded GenAI with a RAG prototype
- Problem
- Learners needed to see why grounding matters.
- What I did
- Used a structured web-data RAG prototype (LangChain prompt chaining, retriever indexing) as a teaching case.
- Outcome
- Grounding taught through a working example.
- Reached people
- Used as a teaching case, so learners saw grounding in a working build.
07 / 11
In the room · how I teach
The other half happens in the room.
-- SQL: group, then totalSELECT region, SUM(sales) AS totalFROM ordersGROUP BY region;
one row per region
// Power BI: a DAX measureTotal Sales = SUM ( Orders[Sales] )Avg Order = DIVIDE ( [Total Sales], COUNTROWS ( Orders ) )
a measure, not a column
# PySpark: the same idea, at scale(df.groupBy("region") .agg(F.sum("sales").alias("total")) .orderBy(F.desc("total")))
nothing runs until an action
// Azure Data Factory: one copy step
one copy step, on a schedule
# GenAI: a prompt with a checkpointprompt: "Summarise this invoice in 3 bullets"draft -> [ human review ] -> send
a person signs off before it goes
# EDA: look at the shape first
look before you model
The kind of example that goes up on the screen.
Who's in the room
- {{a.t}}
How a session runs
- Live coding
- Worked examples from real builds
- Case-based practice
- Facilitator guides
- Walkthroughs and knowledge transfer
Rooms so far
50+trainings and executive education programmes
14+7corporate and academic organisations
Corporate: GlobalFoundries, Honeywell, Freshworks, Air India, Wipro, Aon, RBI, Deloitte, HPE, IBM. Academic: IIT Delhi, IIT Roorkee, IIT Guwahati, JNU, Mangalore University, Manipal Academy of Higher Education.
08 / 11
Expertise
What I bring to the room.
Learning design06
- Training needs analysis
- Curriculum design
- Instructional design
- Learning paths
- Facilitator guides
- Executive education
Data engineering & cloud07
- SQL
- Python
- PySpark
- Azure Data Factory
- Azure Databricks
- ADLS Gen2
- ETL
Analytics & GenAI07
- Power BI
- Tableau
- Generative AI
- LangChain
- RAG
- AI adoption
- Responsible AI
09 / 11
The route · experience, education, certifications
From an IoT lab to the front of the room.
{{sn.kind}}
{{sn.title}}
{{sn.org}}
{{sn.line}}
{{sn.note}}
collected along the way
10 / 11
Questions
Questions?
Write to me about data, GenAI and learning. Thank you for listening.
11 / 11
Presenter notesSlide {{ctr}} / 11
Punitkumar Y Harsur. AI & Data Enablement Consultant, Bengaluru, India. Now Consultant, AI & Machine Learning at Dono Consulting. Before that, Product Strategy & Data Science Consultant at TimesPro. The name assembles; move the pointer through it.
7 years. 50+ corporate trainings and executive education programmes. 3 years of Azure data engineering. 14 corporate and 7 academic organisations. The strings cross at each role: UVCE IoT Lab 2019, UNext 2021, TimesPro 2022, Dono Consulting 2026. Scrub along them, click a ring to pin it, click the figure to replay.
A method, not a menu. Discover: training needs analysis. Design: learning paths and instructional design. Build and deliver: live coding and facilitator guides. Hand over: knowledge transfer. Each press moves one station and ticks the card; the fifth swings the probe out, then the team runs it. The sixth opens the agenda.
Five case studies and one currently designing. Pick a stop on the path, or switch to the list, to read Problem, What I did and Outcome. Cases 01, 02 and 04 have full slides next.
Enterprise AI & Automation Learning Labs. Dono Consulting, 2026. Workflows people build themselves, with approval checkpoints, testing and responsible-AI guidance. Client teams kept running it in their own LMS. It reached people through walkthroughs and knowledge transfer.
Executive education learning paths with IIT partners. TimesPro, 2022 to 2026. Delivered with IIT Delhi, IIT Roorkee and IIT Guwahati. Taught Azure, SQL, Python, PySpark, Power BI and EDA to working professionals.
Teaching grounded GenAI with a RAG prototype. TimesPro. A structured web-data RAG prototype with LangChain prompt chaining and retriever indexing, used as a teaching case.
The teaching side. 50+ corporate trainings and executive education programmes; 14 corporate and 7 academic organisations. Working professionals, executive education participants, early-career learners and mixed-ability business groups. Pick a topic to put a generic worked example on the screen.
Learning design, from training needs analysis to facilitator guides. Data engineering and cloud: SQL, Python, PySpark, Azure Data Factory, Azure Databricks, ADLS Gen2, ETL. Analytics and GenAI: Power BI, Tableau, generative AI, LangChain, RAG, responsible AI. I also write about data engineering.
Experience: UVCE IoT Lab from Mar 2019, UNext Learning from Feb 2021, TimesPro Jun 2022 to Feb 2026, Dono Consulting since Mar 2026. Education: B.E. at RNS Institute of Technology, M.Tech at UVCE, Scaler and TrendyTech. Five certifications. Step with the arrows; the tabs filter the lanes.
Email punitkmr95@gmail.com. LinkedIn linkedin.com/in/punitkmrharsur. GitHub github.com/punitkmryh. Based in Bengaluru, India.
Next up{{upNext}}