How to Build a Sourcing Thesis for Engineering Roles

What is a sourcing thesis in engineering recruitment?

A sourcing thesis is a written plan, agreed before any outreach begins, that says where the right engineers work, what would make them consider a move and how the first message will earn a reply. It replaces keyword searching with a hypothesis you can test and correct. For engineering roles it is the step that decides shortlist quality, because everything after it can only work with the pool it defines.

When an engineering leader asks why their previous recruiter kept sending the wrong profiles, the answer is almost always the same. There was no sourcing thesis. The recruiter knew what the company wanted on paper, ran a keyword search, sent whatever came back, and called it a pipeline. It is not that they were lazy or incompetent. The model they were operating under did not give them the time or the incentive to do anything else.

A sourcing thesis changes this. It is a written document, built before any search begins, that defines the strategy the recruiter will execute. When it is built properly, it is the single most important variable in determining whether a search produces a strong shortlist or a string of near-misses.

What a sourcing thesis is.

A sourcing thesis answers three questions before any outreach begins. The first is where: not "LinkedIn" but specific targets. Which companies employ engineers with the background you need? Which technical communities have they participated in? What signals in someone's public profile or work would identify them as relevant before you even look at their title?

The second question is what attracts them. This is not a list of job benefits. It is the honest, specific case for why this particular opportunity is worth a strong engineer's attention. Senior engineers are not evaluating opportunities in isolation. They are comparing your offer against their current situation, against Razorpay or CRED or Google, against the startup their former colleague just told them about. Your compelling case needs to be built from what is genuinely true about your company and this role, not from marketing copy.

The third question is what will differentiate the outreach. Most senior engineers in India receive several recruitment messages per week. Generic outreach achieves response rates of two to five percent from this group. The sourcing thesis defines the specific angle that will make your outreach worth reading: a reference to the candidate's specific technical work, a precise description of the problem they would be solving, a reason to believe that this conversation is different from the fifteen before it.

Without answers to all three questions, every search is indiscriminate. You are competing for attention in a noisy channel without a strategy for winning that competition.

Why agencies skip this step.

Building a sourcing thesis takes time. For a senior engineering role, doing it properly takes six to ten hours: a deep briefing session with the hiring manager, company and market research to understand the competitive landscape, mapping of the most relevant talent pools, and the actual writing of the document that will guide the search.

Under a contingency model, that time comes entirely out of the agency's margin before any placement fee is confirmed. A contingency recruiter working on ten or fifteen mandates simultaneously cannot afford to invest ten hours in thesis development for each one when fewer than half of those mandates will result in a placement. The incentive is to move fast and submit candidates quickly. Thinking carefully is expensive under that model.

Embedded RPO and dedicated search models fund the thesis-building phase because the fee structure is not contingent on individual placements. The embedded recruiter is paid for time and capacity, not per hire. This removes the structural incentive to skip the strategic work and jump straight to sourcing volume.

The difference in output is not marginal. A recruiter executing against a sourcing thesis consistently delivers shortlists where six out of eight candidates are genuinely relevant. A recruiter executing without one often delivers shortlists where three out of fifteen pass a basic screen. The cost in hiring manager time of the latter scenario compounds quickly across multiple open searches.

Building the "where" component.

The where component of a sourcing thesis starts with company mapping. For a senior backend engineer with experience in high-throughput data systems, the obvious targets are the India engineering teams of companies known to run infrastructure at that scale: Razorpay, Zepto, PhonePe, Swiggy, the India offices of AWS, Google, Atlassian. These are the first-tier targets.

Beyond the obvious tier, the thesis should identify companies that may be contracting, restructuring, or known to have morale challenges. Engineers at companies in that situation are more likely to be open to a conversation than engineers at stable, high-growth environments. This is not about exploiting instability. It is about directing outreach effort toward the pool most likely to be receptive.

Signal-based mapping goes further still. Who is publishing technical content that demonstrates they understand the specific problems you are solving? GitHub activity, conference talks, technical blog posts, and community contributions all surface engineers whose depth is visible before you look at their LinkedIn profile. For RPO in India, this kind of signal-based sourcing consistently outperforms profile-only search because the talent pool is large and the signals that distinguish genuine depth from credential-level familiarity are found in the work, not in the CV.

Building the "what attracts them" component.

Senior engineers evaluate opportunities on a relatively consistent set of dimensions: scope of ownership, technical challenge quality, team calibre, company trajectory, and compensation. The sourcing thesis for a specific role should position the company honestly on each of these dimensions, because the embedded recruiter will use this positioning in every conversation.

Scope of ownership is often the strongest differentiator for product startup and scale-up roles. An engineer who is one of twelve in a team at a large company can often be the technical owner of a significant area at a growth-stage company. That difference in scope matters to engineers who want to grow, and it is a genuine advantage when positioned correctly.

Technical challenge quality matters more than most companies realise. Engineers at the level you are hiring for have seen enough bad codebases and poorly scoped problems to be sceptical of vague descriptions of "exciting challenges." The sourcing thesis should describe the specific technical problem in enough detail that a strong candidate can evaluate whether it is genuinely interesting, not just present it as exciting. Specificity is credibility.

Honesty matters more than spin. A candidate who joins on a premise that turns out to be false leaves quickly, which is the worst outcome for everyone involved. The sourcing thesis should present the company as it actually is, including the genuine challenges, not as the company wishes it were.

Sourcing with a thesis vs sourcing without one.

What makes outreach convert.

Generic outreach that begins with "I came across your profile and thought you might be interested in an exciting opportunity" achieves two to five percent response rates from senior engineering candidates. Thesis-based outreach that opens with a specific reference to the candidate's technical work, names the precise problem they would be working on, and makes a clear and honest case for why this company is worth their attention achieves twenty to forty percent response rates from the same population.

The compounding effect of that difference across a twelve-week search is significant. At five percent response rate with 80 contacts, you are having four conversations. At thirty percent response rate with 80 contacts, you are having 24 conversations. The pipeline that emerges from 24 screened conversations is materially different from the one that emerges from four.

Embedded RPO services in India run this outreach as a sustained programme, not a batch send. Initial contact messages go out in waves. Follow-up messages acknowledge that engineers are busy and keep the door open without being pushy. The relationship is managed over days and weeks, not in a single message. Senior candidates who were not ready to move in week two sometimes become active in week six, and the embedded recruiter is still there to advance those conversations when they mature.

Using the sourcing thesis throughout the search.

The sourcing thesis is not a document written once and filed. It is a living strategy that should be calibrated after each feedback session with the hiring manager. When the first three or four candidates come back with consistent feedback that they are missing something, the right response is to revisit the thesis, not to increase outreach volume. Pipeline health problems are almost always sourcing thesis problems in disguise.

If candidates are technically strong but consistently wrong on the motivation dimension, the "what attracts them" section of the thesis needs refinement. If candidates are motivated but technically weak, the "where" mapping needs to target different companies or different signals. The feedback loop between shortlist response and thesis calibration is the mechanism by which a search improves over time rather than cycling through the same wrong candidates at higher and higher volume.

This calibration is one of the clearest structural advantages of an embedded RPO engagement over an agency relationship. An embedded recruiter attends the feedback sessions, hears the hiring manager's reactions directly, and adjusts the thesis the same day. An agency account manager relays feedback through a layer of translation and the search continues on its existing trajectory for another fortnight.

For a practical guide to the hiring brief that feeds into the sourcing thesis, see how to write an engineering hiring brief that produces better shortlists.

How the models compare: Talhive vs the alternatives.

The table below scores each hiring model against the factors that matter most when building a senior engineering team. Ratings reflect how well each model serves specialist and leadership hiring specifically.

A worked example: an ML engineer who has shipped to production.

The earlier example used a backend role. Here is the same discipline applied to a machine learning engineer, where keyword matching fails badly because so many CVs mention models and few describe running one. Our guide to AI engineer, ML engineer and data scientist roles helps settle which profile you actually need first.

How to tell a sourcing thesis is any good.

You do not need to be a recruiter to review one. Read it before outreach starts and again after the first replies come back.

Where a sourcing thesis fails.

How we write a sourcing thesis.

In an embedded recruitment process outsourcing engagement the thesis is the first written output, and the hiring manager signs it off before outreach starts. It sits alongside the Role Priority Matrix from week one and is revised after each round of shortlist feedback. To see where it fits in the first fortnight, read how embedded RPO works. The brief that feeds it is covered in our guide to writing an engineering hiring brief, and the same approach underpins our engineering and AI hiring.

Frequently asked questions.